2 #include <linux/ceph/ceph_debug.h>
4 #include <linux/module.h>
6 #include <linux/highmem.h>
8 #include <linux/pagemap.h>
9 #include <linux/slab.h>
10 #include <linux/uaccess.h>
12 #include <linux/bio.h>
15 #include <linux/ceph/libceph.h>
16 #include <linux/ceph/osd_client.h>
17 #include <linux/ceph/messenger.h>
18 #include <linux/ceph/decode.h>
19 #include <linux/ceph/auth.h>
20 #include <linux/ceph/pagelist.h>
22 #define OSD_OP_FRONT_LEN 4096
23 #define OSD_OPREPLY_FRONT_LEN 512
25 static struct kmem_cache
*ceph_osd_request_cache
;
27 static const struct ceph_connection_operations osd_con_ops
;
29 static void __send_queued(struct ceph_osd_client
*osdc
);
30 static int __reset_osd(struct ceph_osd_client
*osdc
, struct ceph_osd
*osd
);
31 static void __register_request(struct ceph_osd_client
*osdc
,
32 struct ceph_osd_request
*req
);
33 static void __unregister_request(struct ceph_osd_client
*osdc
,
34 struct ceph_osd_request
*req
);
35 static void __unregister_linger_request(struct ceph_osd_client
*osdc
,
36 struct ceph_osd_request
*req
);
37 static void __enqueue_request(struct ceph_osd_request
*req
);
38 static void __send_request(struct ceph_osd_client
*osdc
,
39 struct ceph_osd_request
*req
);
42 * Implement client access to distributed object storage cluster.
44 * All data objects are stored within a cluster/cloud of OSDs, or
45 * "object storage devices." (Note that Ceph OSDs have _nothing_ to
46 * do with the T10 OSD extensions to SCSI.) Ceph OSDs are simply
47 * remote daemons serving up and coordinating consistent and safe
50 * Cluster membership and the mapping of data objects onto storage devices
51 * are described by the osd map.
53 * We keep track of pending OSD requests (read, write), resubmit
54 * requests to different OSDs when the cluster topology/data layout
55 * change, or retry the affected requests when the communications
56 * channel with an OSD is reset.
60 * calculate the mapping of a file extent onto an object, and fill out the
61 * request accordingly. shorten extent as necessary if it crosses an
64 * fill osd op in request message.
66 static int calc_layout(struct ceph_file_layout
*layout
, u64 off
, u64
*plen
,
67 u64
*objnum
, u64
*objoff
, u64
*objlen
)
73 r
= ceph_calc_file_object_mapping(layout
, off
, orig_len
, objnum
,
77 if (*objlen
< orig_len
) {
79 dout(" skipping last %llu, final file extent %llu~%llu\n",
80 orig_len
- *plen
, off
, *plen
);
83 dout("calc_layout objnum=%llx %llu~%llu\n", *objnum
, *objoff
, *objlen
);
88 static void ceph_osd_data_init(struct ceph_osd_data
*osd_data
)
90 memset(osd_data
, 0, sizeof (*osd_data
));
91 osd_data
->type
= CEPH_OSD_DATA_TYPE_NONE
;
94 static void ceph_osd_data_pages_init(struct ceph_osd_data
*osd_data
,
95 struct page
**pages
, u64 length
, u32 alignment
,
96 bool pages_from_pool
, bool own_pages
)
98 osd_data
->type
= CEPH_OSD_DATA_TYPE_PAGES
;
99 osd_data
->pages
= pages
;
100 osd_data
->length
= length
;
101 osd_data
->alignment
= alignment
;
102 osd_data
->pages_from_pool
= pages_from_pool
;
103 osd_data
->own_pages
= own_pages
;
106 static void ceph_osd_data_pagelist_init(struct ceph_osd_data
*osd_data
,
107 struct ceph_pagelist
*pagelist
)
109 osd_data
->type
= CEPH_OSD_DATA_TYPE_PAGELIST
;
110 osd_data
->pagelist
= pagelist
;
114 static void ceph_osd_data_bio_init(struct ceph_osd_data
*osd_data
,
115 struct bio
*bio
, size_t bio_length
)
117 osd_data
->type
= CEPH_OSD_DATA_TYPE_BIO
;
119 osd_data
->bio_length
= bio_length
;
121 #endif /* CONFIG_BLOCK */
123 #define osd_req_op_data(oreq, whch, typ, fld) \
125 struct ceph_osd_request *__oreq = (oreq); \
126 unsigned int __whch = (whch); \
127 BUG_ON(__whch >= __oreq->r_num_ops); \
128 &__oreq->r_ops[__whch].typ.fld; \
131 static struct ceph_osd_data
*
132 osd_req_op_raw_data_in(struct ceph_osd_request
*osd_req
, unsigned int which
)
134 BUG_ON(which
>= osd_req
->r_num_ops
);
136 return &osd_req
->r_ops
[which
].raw_data_in
;
139 struct ceph_osd_data
*
140 osd_req_op_extent_osd_data(struct ceph_osd_request
*osd_req
,
143 return osd_req_op_data(osd_req
, which
, extent
, osd_data
);
145 EXPORT_SYMBOL(osd_req_op_extent_osd_data
);
147 struct ceph_osd_data
*
148 osd_req_op_cls_response_data(struct ceph_osd_request
*osd_req
,
151 return osd_req_op_data(osd_req
, which
, cls
, response_data
);
153 EXPORT_SYMBOL(osd_req_op_cls_response_data
); /* ??? */
155 void osd_req_op_raw_data_in_pages(struct ceph_osd_request
*osd_req
,
156 unsigned int which
, struct page
**pages
,
157 u64 length
, u32 alignment
,
158 bool pages_from_pool
, bool own_pages
)
160 struct ceph_osd_data
*osd_data
;
162 osd_data
= osd_req_op_raw_data_in(osd_req
, which
);
163 ceph_osd_data_pages_init(osd_data
, pages
, length
, alignment
,
164 pages_from_pool
, own_pages
);
166 EXPORT_SYMBOL(osd_req_op_raw_data_in_pages
);
168 void osd_req_op_extent_osd_data_pages(struct ceph_osd_request
*osd_req
,
169 unsigned int which
, struct page
**pages
,
170 u64 length
, u32 alignment
,
171 bool pages_from_pool
, bool own_pages
)
173 struct ceph_osd_data
*osd_data
;
175 osd_data
= osd_req_op_data(osd_req
, which
, extent
, osd_data
);
176 ceph_osd_data_pages_init(osd_data
, pages
, length
, alignment
,
177 pages_from_pool
, own_pages
);
179 EXPORT_SYMBOL(osd_req_op_extent_osd_data_pages
);
181 void osd_req_op_extent_osd_data_pagelist(struct ceph_osd_request
*osd_req
,
182 unsigned int which
, struct ceph_pagelist
*pagelist
)
184 struct ceph_osd_data
*osd_data
;
186 osd_data
= osd_req_op_data(osd_req
, which
, extent
, osd_data
);
187 ceph_osd_data_pagelist_init(osd_data
, pagelist
);
189 EXPORT_SYMBOL(osd_req_op_extent_osd_data_pagelist
);
192 void osd_req_op_extent_osd_data_bio(struct ceph_osd_request
*osd_req
,
193 unsigned int which
, struct bio
*bio
, size_t bio_length
)
195 struct ceph_osd_data
*osd_data
;
197 osd_data
= osd_req_op_data(osd_req
, which
, extent
, osd_data
);
198 ceph_osd_data_bio_init(osd_data
, bio
, bio_length
);
200 EXPORT_SYMBOL(osd_req_op_extent_osd_data_bio
);
201 #endif /* CONFIG_BLOCK */
203 static void osd_req_op_cls_request_info_pagelist(
204 struct ceph_osd_request
*osd_req
,
205 unsigned int which
, struct ceph_pagelist
*pagelist
)
207 struct ceph_osd_data
*osd_data
;
209 osd_data
= osd_req_op_data(osd_req
, which
, cls
, request_info
);
210 ceph_osd_data_pagelist_init(osd_data
, pagelist
);
213 void osd_req_op_cls_request_data_pagelist(
214 struct ceph_osd_request
*osd_req
,
215 unsigned int which
, struct ceph_pagelist
*pagelist
)
217 struct ceph_osd_data
*osd_data
;
219 osd_data
= osd_req_op_data(osd_req
, which
, cls
, request_data
);
220 ceph_osd_data_pagelist_init(osd_data
, pagelist
);
222 EXPORT_SYMBOL(osd_req_op_cls_request_data_pagelist
);
224 void osd_req_op_cls_request_data_pages(struct ceph_osd_request
*osd_req
,
225 unsigned int which
, struct page
**pages
, u64 length
,
226 u32 alignment
, bool pages_from_pool
, bool own_pages
)
228 struct ceph_osd_data
*osd_data
;
230 osd_data
= osd_req_op_data(osd_req
, which
, cls
, request_data
);
231 ceph_osd_data_pages_init(osd_data
, pages
, length
, alignment
,
232 pages_from_pool
, own_pages
);
234 EXPORT_SYMBOL(osd_req_op_cls_request_data_pages
);
236 void osd_req_op_cls_response_data_pages(struct ceph_osd_request
*osd_req
,
237 unsigned int which
, struct page
**pages
, u64 length
,
238 u32 alignment
, bool pages_from_pool
, bool own_pages
)
240 struct ceph_osd_data
*osd_data
;
242 osd_data
= osd_req_op_data(osd_req
, which
, cls
, response_data
);
243 ceph_osd_data_pages_init(osd_data
, pages
, length
, alignment
,
244 pages_from_pool
, own_pages
);
246 EXPORT_SYMBOL(osd_req_op_cls_response_data_pages
);
248 static u64
ceph_osd_data_length(struct ceph_osd_data
*osd_data
)
250 switch (osd_data
->type
) {
251 case CEPH_OSD_DATA_TYPE_NONE
:
253 case CEPH_OSD_DATA_TYPE_PAGES
:
254 return osd_data
->length
;
255 case CEPH_OSD_DATA_TYPE_PAGELIST
:
256 return (u64
)osd_data
->pagelist
->length
;
258 case CEPH_OSD_DATA_TYPE_BIO
:
259 return (u64
)osd_data
->bio_length
;
260 #endif /* CONFIG_BLOCK */
262 WARN(true, "unrecognized data type %d\n", (int)osd_data
->type
);
267 static void ceph_osd_data_release(struct ceph_osd_data
*osd_data
)
269 if (osd_data
->type
== CEPH_OSD_DATA_TYPE_PAGES
&& osd_data
->own_pages
) {
272 num_pages
= calc_pages_for((u64
)osd_data
->alignment
,
273 (u64
)osd_data
->length
);
274 ceph_release_page_vector(osd_data
->pages
, num_pages
);
276 ceph_osd_data_init(osd_data
);
279 static void osd_req_op_data_release(struct ceph_osd_request
*osd_req
,
282 struct ceph_osd_req_op
*op
;
284 BUG_ON(which
>= osd_req
->r_num_ops
);
285 op
= &osd_req
->r_ops
[which
];
288 case CEPH_OSD_OP_READ
:
289 case CEPH_OSD_OP_WRITE
:
290 case CEPH_OSD_OP_WRITEFULL
:
291 ceph_osd_data_release(&op
->extent
.osd_data
);
293 case CEPH_OSD_OP_CALL
:
294 ceph_osd_data_release(&op
->cls
.request_info
);
295 ceph_osd_data_release(&op
->cls
.request_data
);
296 ceph_osd_data_release(&op
->cls
.response_data
);
298 case CEPH_OSD_OP_SETXATTR
:
299 case CEPH_OSD_OP_CMPXATTR
:
300 ceph_osd_data_release(&op
->xattr
.osd_data
);
302 case CEPH_OSD_OP_STAT
:
303 ceph_osd_data_release(&op
->raw_data_in
);
313 static void ceph_osdc_release_request(struct kref
*kref
)
315 struct ceph_osd_request
*req
= container_of(kref
,
316 struct ceph_osd_request
, r_kref
);
319 dout("%s %p (r_request %p r_reply %p)\n", __func__
, req
,
320 req
->r_request
, req
->r_reply
);
321 WARN_ON(!RB_EMPTY_NODE(&req
->r_node
));
322 WARN_ON(!list_empty(&req
->r_req_lru_item
));
323 WARN_ON(!list_empty(&req
->r_osd_item
));
324 WARN_ON(!list_empty(&req
->r_linger_item
));
325 WARN_ON(!list_empty(&req
->r_linger_osd_item
));
329 ceph_msg_put(req
->r_request
);
331 ceph_msg_revoke_incoming(req
->r_reply
);
332 ceph_msg_put(req
->r_reply
);
335 for (which
= 0; which
< req
->r_num_ops
; which
++)
336 osd_req_op_data_release(req
, which
);
338 ceph_put_snap_context(req
->r_snapc
);
340 mempool_free(req
, req
->r_osdc
->req_mempool
);
341 else if (req
->r_num_ops
<= CEPH_OSD_SLAB_OPS
)
342 kmem_cache_free(ceph_osd_request_cache
, req
);
347 void ceph_osdc_get_request(struct ceph_osd_request
*req
)
349 dout("%s %p (was %d)\n", __func__
, req
,
350 atomic_read(&req
->r_kref
.refcount
));
351 kref_get(&req
->r_kref
);
353 EXPORT_SYMBOL(ceph_osdc_get_request
);
355 void ceph_osdc_put_request(struct ceph_osd_request
*req
)
358 dout("%s %p (was %d)\n", __func__
, req
,
359 atomic_read(&req
->r_kref
.refcount
));
360 kref_put(&req
->r_kref
, ceph_osdc_release_request
);
363 EXPORT_SYMBOL(ceph_osdc_put_request
);
365 struct ceph_osd_request
*ceph_osdc_alloc_request(struct ceph_osd_client
*osdc
,
366 struct ceph_snap_context
*snapc
,
367 unsigned int num_ops
,
371 struct ceph_osd_request
*req
;
372 struct ceph_msg
*msg
;
376 BUG_ON(num_ops
> CEPH_OSD_SLAB_OPS
);
377 req
= mempool_alloc(osdc
->req_mempool
, gfp_flags
);
378 } else if (num_ops
<= CEPH_OSD_SLAB_OPS
) {
379 req
= kmem_cache_alloc(ceph_osd_request_cache
, gfp_flags
);
381 BUG_ON(num_ops
> CEPH_OSD_MAX_OPS
);
382 req
= kmalloc(sizeof(*req
) + num_ops
* sizeof(req
->r_ops
[0]),
388 /* req only, each op is zeroed in _osd_req_op_init() */
389 memset(req
, 0, sizeof(*req
));
392 req
->r_mempool
= use_mempool
;
393 req
->r_num_ops
= num_ops
;
394 req
->r_snapid
= CEPH_NOSNAP
;
395 req
->r_snapc
= ceph_get_snap_context(snapc
);
397 kref_init(&req
->r_kref
);
398 init_completion(&req
->r_completion
);
399 init_completion(&req
->r_safe_completion
);
400 RB_CLEAR_NODE(&req
->r_node
);
401 INIT_LIST_HEAD(&req
->r_unsafe_item
);
402 INIT_LIST_HEAD(&req
->r_linger_item
);
403 INIT_LIST_HEAD(&req
->r_linger_osd_item
);
404 INIT_LIST_HEAD(&req
->r_req_lru_item
);
405 INIT_LIST_HEAD(&req
->r_osd_item
);
407 req
->r_base_oloc
.pool
= -1;
408 req
->r_target_oloc
.pool
= -1;
410 msg_size
= OSD_OPREPLY_FRONT_LEN
;
411 if (num_ops
> CEPH_OSD_SLAB_OPS
) {
412 /* ceph_osd_op and rval */
413 msg_size
+= (num_ops
- CEPH_OSD_SLAB_OPS
) *
414 (sizeof(struct ceph_osd_op
) + 4);
417 /* create reply message */
419 msg
= ceph_msgpool_get(&osdc
->msgpool_op_reply
, 0);
421 msg
= ceph_msg_new(CEPH_MSG_OSD_OPREPLY
, msg_size
,
424 ceph_osdc_put_request(req
);
429 msg_size
= 4 + 4 + 4; /* client_inc, osdmap_epoch, flags */
430 msg_size
+= 4 + 4 + 4 + 8; /* mtime, reassert_version */
431 msg_size
+= 2 + 4 + 8 + 4 + 4; /* oloc */
432 msg_size
+= 1 + 8 + 4 + 4; /* pgid */
433 msg_size
+= 4 + CEPH_MAX_OID_NAME_LEN
; /* oid */
434 msg_size
+= 2 + num_ops
* sizeof(struct ceph_osd_op
);
435 msg_size
+= 8; /* snapid */
436 msg_size
+= 8; /* snap_seq */
437 msg_size
+= 4 + 8 * (snapc
? snapc
->num_snaps
: 0); /* snaps */
438 msg_size
+= 4; /* retry_attempt */
440 /* create request message; allow space for oid */
442 msg
= ceph_msgpool_get(&osdc
->msgpool_op
, 0);
444 msg
= ceph_msg_new(CEPH_MSG_OSD_OP
, msg_size
, gfp_flags
, true);
446 ceph_osdc_put_request(req
);
450 memset(msg
->front
.iov_base
, 0, msg
->front
.iov_len
);
452 req
->r_request
= msg
;
456 EXPORT_SYMBOL(ceph_osdc_alloc_request
);
458 static bool osd_req_opcode_valid(u16 opcode
)
461 #define GENERATE_CASE(op, opcode, str) case CEPH_OSD_OP_##op: return true;
462 __CEPH_FORALL_OSD_OPS(GENERATE_CASE
)
470 * This is an osd op init function for opcodes that have no data or
471 * other information associated with them. It also serves as a
472 * common init routine for all the other init functions, below.
474 static struct ceph_osd_req_op
*
475 _osd_req_op_init(struct ceph_osd_request
*osd_req
, unsigned int which
,
476 u16 opcode
, u32 flags
)
478 struct ceph_osd_req_op
*op
;
480 BUG_ON(which
>= osd_req
->r_num_ops
);
481 BUG_ON(!osd_req_opcode_valid(opcode
));
483 op
= &osd_req
->r_ops
[which
];
484 memset(op
, 0, sizeof (*op
));
491 void osd_req_op_init(struct ceph_osd_request
*osd_req
,
492 unsigned int which
, u16 opcode
, u32 flags
)
494 (void)_osd_req_op_init(osd_req
, which
, opcode
, flags
);
496 EXPORT_SYMBOL(osd_req_op_init
);
498 void osd_req_op_extent_init(struct ceph_osd_request
*osd_req
,
499 unsigned int which
, u16 opcode
,
500 u64 offset
, u64 length
,
501 u64 truncate_size
, u32 truncate_seq
)
503 struct ceph_osd_req_op
*op
= _osd_req_op_init(osd_req
, which
,
505 size_t payload_len
= 0;
507 BUG_ON(opcode
!= CEPH_OSD_OP_READ
&& opcode
!= CEPH_OSD_OP_WRITE
&&
508 opcode
!= CEPH_OSD_OP_WRITEFULL
&& opcode
!= CEPH_OSD_OP_ZERO
&&
509 opcode
!= CEPH_OSD_OP_TRUNCATE
);
511 op
->extent
.offset
= offset
;
512 op
->extent
.length
= length
;
513 op
->extent
.truncate_size
= truncate_size
;
514 op
->extent
.truncate_seq
= truncate_seq
;
515 if (opcode
== CEPH_OSD_OP_WRITE
|| opcode
== CEPH_OSD_OP_WRITEFULL
)
516 payload_len
+= length
;
518 op
->indata_len
= payload_len
;
520 EXPORT_SYMBOL(osd_req_op_extent_init
);
522 void osd_req_op_extent_update(struct ceph_osd_request
*osd_req
,
523 unsigned int which
, u64 length
)
525 struct ceph_osd_req_op
*op
;
528 BUG_ON(which
>= osd_req
->r_num_ops
);
529 op
= &osd_req
->r_ops
[which
];
530 previous
= op
->extent
.length
;
532 if (length
== previous
)
533 return; /* Nothing to do */
534 BUG_ON(length
> previous
);
536 op
->extent
.length
= length
;
537 op
->indata_len
-= previous
- length
;
539 EXPORT_SYMBOL(osd_req_op_extent_update
);
541 void osd_req_op_extent_dup_last(struct ceph_osd_request
*osd_req
,
542 unsigned int which
, u64 offset_inc
)
544 struct ceph_osd_req_op
*op
, *prev_op
;
546 BUG_ON(which
+ 1 >= osd_req
->r_num_ops
);
548 prev_op
= &osd_req
->r_ops
[which
];
549 op
= _osd_req_op_init(osd_req
, which
+ 1, prev_op
->op
, prev_op
->flags
);
550 /* dup previous one */
551 op
->indata_len
= prev_op
->indata_len
;
552 op
->outdata_len
= prev_op
->outdata_len
;
553 op
->extent
= prev_op
->extent
;
555 op
->extent
.offset
+= offset_inc
;
556 op
->extent
.length
-= offset_inc
;
558 if (op
->op
== CEPH_OSD_OP_WRITE
|| op
->op
== CEPH_OSD_OP_WRITEFULL
)
559 op
->indata_len
-= offset_inc
;
561 EXPORT_SYMBOL(osd_req_op_extent_dup_last
);
563 void osd_req_op_cls_init(struct ceph_osd_request
*osd_req
, unsigned int which
,
564 u16 opcode
, const char *class, const char *method
)
566 struct ceph_osd_req_op
*op
= _osd_req_op_init(osd_req
, which
,
568 struct ceph_pagelist
*pagelist
;
569 size_t payload_len
= 0;
572 BUG_ON(opcode
!= CEPH_OSD_OP_CALL
);
574 pagelist
= kmalloc(sizeof (*pagelist
), GFP_NOFS
);
576 ceph_pagelist_init(pagelist
);
578 op
->cls
.class_name
= class;
579 size
= strlen(class);
580 BUG_ON(size
> (size_t) U8_MAX
);
581 op
->cls
.class_len
= size
;
582 ceph_pagelist_append(pagelist
, class, size
);
585 op
->cls
.method_name
= method
;
586 size
= strlen(method
);
587 BUG_ON(size
> (size_t) U8_MAX
);
588 op
->cls
.method_len
= size
;
589 ceph_pagelist_append(pagelist
, method
, size
);
592 osd_req_op_cls_request_info_pagelist(osd_req
, which
, pagelist
);
594 op
->cls
.argc
= 0; /* currently unused */
596 op
->indata_len
= payload_len
;
598 EXPORT_SYMBOL(osd_req_op_cls_init
);
600 int osd_req_op_xattr_init(struct ceph_osd_request
*osd_req
, unsigned int which
,
601 u16 opcode
, const char *name
, const void *value
,
602 size_t size
, u8 cmp_op
, u8 cmp_mode
)
604 struct ceph_osd_req_op
*op
= _osd_req_op_init(osd_req
, which
,
606 struct ceph_pagelist
*pagelist
;
609 BUG_ON(opcode
!= CEPH_OSD_OP_SETXATTR
&& opcode
!= CEPH_OSD_OP_CMPXATTR
);
611 pagelist
= kmalloc(sizeof(*pagelist
), GFP_NOFS
);
615 ceph_pagelist_init(pagelist
);
617 payload_len
= strlen(name
);
618 op
->xattr
.name_len
= payload_len
;
619 ceph_pagelist_append(pagelist
, name
, payload_len
);
621 op
->xattr
.value_len
= size
;
622 ceph_pagelist_append(pagelist
, value
, size
);
625 op
->xattr
.cmp_op
= cmp_op
;
626 op
->xattr
.cmp_mode
= cmp_mode
;
628 ceph_osd_data_pagelist_init(&op
->xattr
.osd_data
, pagelist
);
629 op
->indata_len
= payload_len
;
632 EXPORT_SYMBOL(osd_req_op_xattr_init
);
634 void osd_req_op_watch_init(struct ceph_osd_request
*osd_req
,
635 unsigned int which
, u16 opcode
,
636 u64 cookie
, u64 version
, int flag
)
638 struct ceph_osd_req_op
*op
= _osd_req_op_init(osd_req
, which
,
641 BUG_ON(opcode
!= CEPH_OSD_OP_NOTIFY_ACK
&& opcode
!= CEPH_OSD_OP_WATCH
);
643 op
->watch
.cookie
= cookie
;
644 op
->watch
.ver
= version
;
645 if (opcode
== CEPH_OSD_OP_WATCH
&& flag
)
646 op
->watch
.flag
= (u8
)1;
648 EXPORT_SYMBOL(osd_req_op_watch_init
);
650 void osd_req_op_alloc_hint_init(struct ceph_osd_request
*osd_req
,
652 u64 expected_object_size
,
653 u64 expected_write_size
)
655 struct ceph_osd_req_op
*op
= _osd_req_op_init(osd_req
, which
,
656 CEPH_OSD_OP_SETALLOCHINT
,
659 op
->alloc_hint
.expected_object_size
= expected_object_size
;
660 op
->alloc_hint
.expected_write_size
= expected_write_size
;
663 * CEPH_OSD_OP_SETALLOCHINT op is advisory and therefore deemed
664 * not worth a feature bit. Set FAILOK per-op flag to make
665 * sure older osds don't trip over an unsupported opcode.
667 op
->flags
|= CEPH_OSD_OP_FLAG_FAILOK
;
669 EXPORT_SYMBOL(osd_req_op_alloc_hint_init
);
671 static void ceph_osdc_msg_data_add(struct ceph_msg
*msg
,
672 struct ceph_osd_data
*osd_data
)
674 u64 length
= ceph_osd_data_length(osd_data
);
676 if (osd_data
->type
== CEPH_OSD_DATA_TYPE_PAGES
) {
677 BUG_ON(length
> (u64
) SIZE_MAX
);
679 ceph_msg_data_add_pages(msg
, osd_data
->pages
,
680 length
, osd_data
->alignment
);
681 } else if (osd_data
->type
== CEPH_OSD_DATA_TYPE_PAGELIST
) {
683 ceph_msg_data_add_pagelist(msg
, osd_data
->pagelist
);
685 } else if (osd_data
->type
== CEPH_OSD_DATA_TYPE_BIO
) {
686 ceph_msg_data_add_bio(msg
, osd_data
->bio
, length
);
689 BUG_ON(osd_data
->type
!= CEPH_OSD_DATA_TYPE_NONE
);
693 static u64
osd_req_encode_op(struct ceph_osd_request
*req
,
694 struct ceph_osd_op
*dst
, unsigned int which
)
696 struct ceph_osd_req_op
*src
;
697 struct ceph_osd_data
*osd_data
;
698 u64 request_data_len
= 0;
701 BUG_ON(which
>= req
->r_num_ops
);
702 src
= &req
->r_ops
[which
];
703 if (WARN_ON(!osd_req_opcode_valid(src
->op
))) {
704 pr_err("unrecognized osd opcode %d\n", src
->op
);
710 case CEPH_OSD_OP_STAT
:
711 osd_data
= &src
->raw_data_in
;
712 ceph_osdc_msg_data_add(req
->r_reply
, osd_data
);
714 case CEPH_OSD_OP_READ
:
715 case CEPH_OSD_OP_WRITE
:
716 case CEPH_OSD_OP_WRITEFULL
:
717 case CEPH_OSD_OP_ZERO
:
718 case CEPH_OSD_OP_TRUNCATE
:
719 if (src
->op
== CEPH_OSD_OP_WRITE
||
720 src
->op
== CEPH_OSD_OP_WRITEFULL
)
721 request_data_len
= src
->extent
.length
;
722 dst
->extent
.offset
= cpu_to_le64(src
->extent
.offset
);
723 dst
->extent
.length
= cpu_to_le64(src
->extent
.length
);
724 dst
->extent
.truncate_size
=
725 cpu_to_le64(src
->extent
.truncate_size
);
726 dst
->extent
.truncate_seq
=
727 cpu_to_le32(src
->extent
.truncate_seq
);
728 osd_data
= &src
->extent
.osd_data
;
729 if (src
->op
== CEPH_OSD_OP_WRITE
||
730 src
->op
== CEPH_OSD_OP_WRITEFULL
)
731 ceph_osdc_msg_data_add(req
->r_request
, osd_data
);
733 ceph_osdc_msg_data_add(req
->r_reply
, osd_data
);
735 case CEPH_OSD_OP_CALL
:
736 dst
->cls
.class_len
= src
->cls
.class_len
;
737 dst
->cls
.method_len
= src
->cls
.method_len
;
738 osd_data
= &src
->cls
.request_info
;
739 ceph_osdc_msg_data_add(req
->r_request
, osd_data
);
740 BUG_ON(osd_data
->type
!= CEPH_OSD_DATA_TYPE_PAGELIST
);
741 request_data_len
= osd_data
->pagelist
->length
;
743 osd_data
= &src
->cls
.request_data
;
744 data_length
= ceph_osd_data_length(osd_data
);
746 BUG_ON(osd_data
->type
== CEPH_OSD_DATA_TYPE_NONE
);
747 dst
->cls
.indata_len
= cpu_to_le32(data_length
);
748 ceph_osdc_msg_data_add(req
->r_request
, osd_data
);
749 src
->indata_len
+= data_length
;
750 request_data_len
+= data_length
;
752 osd_data
= &src
->cls
.response_data
;
753 ceph_osdc_msg_data_add(req
->r_reply
, osd_data
);
755 case CEPH_OSD_OP_STARTSYNC
:
757 case CEPH_OSD_OP_NOTIFY_ACK
:
758 case CEPH_OSD_OP_WATCH
:
759 dst
->watch
.cookie
= cpu_to_le64(src
->watch
.cookie
);
760 dst
->watch
.ver
= cpu_to_le64(src
->watch
.ver
);
761 dst
->watch
.flag
= src
->watch
.flag
;
763 case CEPH_OSD_OP_SETALLOCHINT
:
764 dst
->alloc_hint
.expected_object_size
=
765 cpu_to_le64(src
->alloc_hint
.expected_object_size
);
766 dst
->alloc_hint
.expected_write_size
=
767 cpu_to_le64(src
->alloc_hint
.expected_write_size
);
769 case CEPH_OSD_OP_SETXATTR
:
770 case CEPH_OSD_OP_CMPXATTR
:
771 dst
->xattr
.name_len
= cpu_to_le32(src
->xattr
.name_len
);
772 dst
->xattr
.value_len
= cpu_to_le32(src
->xattr
.value_len
);
773 dst
->xattr
.cmp_op
= src
->xattr
.cmp_op
;
774 dst
->xattr
.cmp_mode
= src
->xattr
.cmp_mode
;
775 osd_data
= &src
->xattr
.osd_data
;
776 ceph_osdc_msg_data_add(req
->r_request
, osd_data
);
777 request_data_len
= osd_data
->pagelist
->length
;
779 case CEPH_OSD_OP_CREATE
:
780 case CEPH_OSD_OP_DELETE
:
783 pr_err("unsupported osd opcode %s\n",
784 ceph_osd_op_name(src
->op
));
790 dst
->op
= cpu_to_le16(src
->op
);
791 dst
->flags
= cpu_to_le32(src
->flags
);
792 dst
->payload_len
= cpu_to_le32(src
->indata_len
);
794 return request_data_len
;
798 * build new request AND message, calculate layout, and adjust file
801 * if the file was recently truncated, we include information about its
802 * old and new size so that the object can be updated appropriately. (we
803 * avoid synchronously deleting truncated objects because it's slow.)
805 * if @do_sync, include a 'startsync' command so that the osd will flush
808 struct ceph_osd_request
*ceph_osdc_new_request(struct ceph_osd_client
*osdc
,
809 struct ceph_file_layout
*layout
,
810 struct ceph_vino vino
,
812 unsigned int which
, int num_ops
,
813 int opcode
, int flags
,
814 struct ceph_snap_context
*snapc
,
819 struct ceph_osd_request
*req
;
825 BUG_ON(opcode
!= CEPH_OSD_OP_READ
&& opcode
!= CEPH_OSD_OP_WRITE
&&
826 opcode
!= CEPH_OSD_OP_ZERO
&& opcode
!= CEPH_OSD_OP_TRUNCATE
&&
827 opcode
!= CEPH_OSD_OP_CREATE
&& opcode
!= CEPH_OSD_OP_DELETE
);
829 req
= ceph_osdc_alloc_request(osdc
, snapc
, num_ops
, use_mempool
,
832 return ERR_PTR(-ENOMEM
);
834 req
->r_flags
= flags
;
836 /* calculate max write size */
837 r
= calc_layout(layout
, off
, plen
, &objnum
, &objoff
, &objlen
);
839 ceph_osdc_put_request(req
);
843 if (opcode
== CEPH_OSD_OP_CREATE
|| opcode
== CEPH_OSD_OP_DELETE
) {
844 osd_req_op_init(req
, which
, opcode
, 0);
846 u32 object_size
= le32_to_cpu(layout
->fl_object_size
);
847 u32 object_base
= off
- objoff
;
848 if (!(truncate_seq
== 1 && truncate_size
== -1ULL)) {
849 if (truncate_size
<= object_base
) {
852 truncate_size
-= object_base
;
853 if (truncate_size
> object_size
)
854 truncate_size
= object_size
;
857 osd_req_op_extent_init(req
, which
, opcode
, objoff
, objlen
,
858 truncate_size
, truncate_seq
);
861 req
->r_base_oloc
.pool
= ceph_file_layout_pg_pool(*layout
);
863 snprintf(req
->r_base_oid
.name
, sizeof(req
->r_base_oid
.name
),
864 "%llx.%08llx", vino
.ino
, objnum
);
865 req
->r_base_oid
.name_len
= strlen(req
->r_base_oid
.name
);
869 EXPORT_SYMBOL(ceph_osdc_new_request
);
872 * We keep osd requests in an rbtree, sorted by ->r_tid.
874 static void __insert_request(struct ceph_osd_client
*osdc
,
875 struct ceph_osd_request
*new)
877 struct rb_node
**p
= &osdc
->requests
.rb_node
;
878 struct rb_node
*parent
= NULL
;
879 struct ceph_osd_request
*req
= NULL
;
883 req
= rb_entry(parent
, struct ceph_osd_request
, r_node
);
884 if (new->r_tid
< req
->r_tid
)
886 else if (new->r_tid
> req
->r_tid
)
892 rb_link_node(&new->r_node
, parent
, p
);
893 rb_insert_color(&new->r_node
, &osdc
->requests
);
896 static struct ceph_osd_request
*__lookup_request(struct ceph_osd_client
*osdc
,
899 struct ceph_osd_request
*req
;
900 struct rb_node
*n
= osdc
->requests
.rb_node
;
903 req
= rb_entry(n
, struct ceph_osd_request
, r_node
);
904 if (tid
< req
->r_tid
)
906 else if (tid
> req
->r_tid
)
914 static struct ceph_osd_request
*
915 __lookup_request_ge(struct ceph_osd_client
*osdc
,
918 struct ceph_osd_request
*req
;
919 struct rb_node
*n
= osdc
->requests
.rb_node
;
922 req
= rb_entry(n
, struct ceph_osd_request
, r_node
);
923 if (tid
< req
->r_tid
) {
927 } else if (tid
> req
->r_tid
) {
936 static void __kick_linger_request(struct ceph_osd_request
*req
)
938 struct ceph_osd_client
*osdc
= req
->r_osdc
;
939 struct ceph_osd
*osd
= req
->r_osd
;
942 * Linger requests need to be resent with a new tid to avoid
943 * the dup op detection logic on the OSDs. Achieve this with
944 * a re-register dance instead of open-coding.
946 ceph_osdc_get_request(req
);
947 if (!list_empty(&req
->r_linger_item
))
948 __unregister_linger_request(osdc
, req
);
950 __unregister_request(osdc
, req
);
951 __register_request(osdc
, req
);
952 ceph_osdc_put_request(req
);
955 * Unless request has been registered as both normal and
956 * lingering, __unregister{,_linger}_request clears r_osd.
957 * However, here we need to preserve r_osd to make sure we
958 * requeue on the same OSD.
960 WARN_ON(req
->r_osd
|| !osd
);
963 dout("%s requeueing %p tid %llu\n", __func__
, req
, req
->r_tid
);
964 __enqueue_request(req
);
968 * Resubmit requests pending on the given osd.
970 static void __kick_osd_requests(struct ceph_osd_client
*osdc
,
971 struct ceph_osd
*osd
)
973 struct ceph_osd_request
*req
, *nreq
;
975 LIST_HEAD(resend_linger
);
978 dout("%s osd%d\n", __func__
, osd
->o_osd
);
979 err
= __reset_osd(osdc
, osd
);
984 * Build up a list of requests to resend by traversing the
985 * osd's list of requests. Requests for a given object are
986 * sent in tid order, and that is also the order they're
987 * kept on this list. Therefore all requests that are in
988 * flight will be found first, followed by all requests that
989 * have not yet been sent. And to resend requests while
990 * preserving this order we will want to put any sent
991 * requests back on the front of the osd client's unsent
994 * So we build a separate ordered list of already-sent
995 * requests for the affected osd and splice it onto the
996 * front of the osd client's unsent list. Once we've seen a
997 * request that has not yet been sent we're done. Those
998 * requests are already sitting right where they belong.
1000 list_for_each_entry(req
, &osd
->o_requests
, r_osd_item
) {
1004 if (!req
->r_linger
) {
1005 dout("%s requeueing %p tid %llu\n", __func__
, req
,
1007 list_move_tail(&req
->r_req_lru_item
, &resend
);
1008 req
->r_flags
|= CEPH_OSD_FLAG_RETRY
;
1010 list_move_tail(&req
->r_req_lru_item
, &resend_linger
);
1013 list_splice(&resend
, &osdc
->req_unsent
);
1016 * Both registered and not yet registered linger requests are
1017 * enqueued with a new tid on the same OSD. We add/move them
1018 * to req_unsent/o_requests at the end to keep things in tid
1021 list_for_each_entry_safe(req
, nreq
, &osd
->o_linger_requests
,
1022 r_linger_osd_item
) {
1023 WARN_ON(!list_empty(&req
->r_req_lru_item
));
1024 __kick_linger_request(req
);
1027 list_for_each_entry_safe(req
, nreq
, &resend_linger
, r_req_lru_item
)
1028 __kick_linger_request(req
);
1032 * If the osd connection drops, we need to resubmit all requests.
1034 static void osd_reset(struct ceph_connection
*con
)
1036 struct ceph_osd
*osd
= con
->private;
1037 struct ceph_osd_client
*osdc
;
1041 dout("osd_reset osd%d\n", osd
->o_osd
);
1043 down_read(&osdc
->map_sem
);
1044 mutex_lock(&osdc
->request_mutex
);
1045 __kick_osd_requests(osdc
, osd
);
1046 __send_queued(osdc
);
1047 mutex_unlock(&osdc
->request_mutex
);
1048 up_read(&osdc
->map_sem
);
1052 * Track open sessions with osds.
1054 static struct ceph_osd
*create_osd(struct ceph_osd_client
*osdc
, int onum
)
1056 struct ceph_osd
*osd
;
1058 osd
= kzalloc(sizeof(*osd
), GFP_NOFS
);
1062 atomic_set(&osd
->o_ref
, 1);
1065 RB_CLEAR_NODE(&osd
->o_node
);
1066 INIT_LIST_HEAD(&osd
->o_requests
);
1067 INIT_LIST_HEAD(&osd
->o_linger_requests
);
1068 INIT_LIST_HEAD(&osd
->o_osd_lru
);
1069 osd
->o_incarnation
= 1;
1071 ceph_con_init(&osd
->o_con
, osd
, &osd_con_ops
, &osdc
->client
->msgr
);
1073 INIT_LIST_HEAD(&osd
->o_keepalive_item
);
1077 static struct ceph_osd
*get_osd(struct ceph_osd
*osd
)
1079 if (atomic_inc_not_zero(&osd
->o_ref
)) {
1080 dout("get_osd %p %d -> %d\n", osd
, atomic_read(&osd
->o_ref
)-1,
1081 atomic_read(&osd
->o_ref
));
1084 dout("get_osd %p FAIL\n", osd
);
1089 static void put_osd(struct ceph_osd
*osd
)
1091 dout("put_osd %p %d -> %d\n", osd
, atomic_read(&osd
->o_ref
),
1092 atomic_read(&osd
->o_ref
) - 1);
1093 if (atomic_dec_and_test(&osd
->o_ref
)) {
1094 if (osd
->o_auth
.authorizer
)
1095 ceph_auth_destroy_authorizer(osd
->o_auth
.authorizer
);
1101 * remove an osd from our map
1103 static void __remove_osd(struct ceph_osd_client
*osdc
, struct ceph_osd
*osd
)
1105 dout("%s %p osd%d\n", __func__
, osd
, osd
->o_osd
);
1106 WARN_ON(!list_empty(&osd
->o_requests
));
1107 WARN_ON(!list_empty(&osd
->o_linger_requests
));
1109 list_del_init(&osd
->o_osd_lru
);
1110 rb_erase(&osd
->o_node
, &osdc
->osds
);
1111 RB_CLEAR_NODE(&osd
->o_node
);
1114 static void remove_osd(struct ceph_osd_client
*osdc
, struct ceph_osd
*osd
)
1116 dout("%s %p osd%d\n", __func__
, osd
, osd
->o_osd
);
1118 if (!RB_EMPTY_NODE(&osd
->o_node
)) {
1119 ceph_con_close(&osd
->o_con
);
1120 __remove_osd(osdc
, osd
);
1125 static void remove_all_osds(struct ceph_osd_client
*osdc
)
1127 dout("%s %p\n", __func__
, osdc
);
1128 mutex_lock(&osdc
->request_mutex
);
1129 while (!RB_EMPTY_ROOT(&osdc
->osds
)) {
1130 struct ceph_osd
*osd
= rb_entry(rb_first(&osdc
->osds
),
1131 struct ceph_osd
, o_node
);
1132 remove_osd(osdc
, osd
);
1134 mutex_unlock(&osdc
->request_mutex
);
1137 static void __move_osd_to_lru(struct ceph_osd_client
*osdc
,
1138 struct ceph_osd
*osd
)
1140 dout("%s %p\n", __func__
, osd
);
1141 BUG_ON(!list_empty(&osd
->o_osd_lru
));
1143 list_add_tail(&osd
->o_osd_lru
, &osdc
->osd_lru
);
1144 osd
->lru_ttl
= jiffies
+ osdc
->client
->options
->osd_idle_ttl
;
1147 static void maybe_move_osd_to_lru(struct ceph_osd_client
*osdc
,
1148 struct ceph_osd
*osd
)
1150 dout("%s %p\n", __func__
, osd
);
1152 if (list_empty(&osd
->o_requests
) &&
1153 list_empty(&osd
->o_linger_requests
))
1154 __move_osd_to_lru(osdc
, osd
);
1157 static void __remove_osd_from_lru(struct ceph_osd
*osd
)
1159 dout("__remove_osd_from_lru %p\n", osd
);
1160 if (!list_empty(&osd
->o_osd_lru
))
1161 list_del_init(&osd
->o_osd_lru
);
1164 static void remove_old_osds(struct ceph_osd_client
*osdc
)
1166 struct ceph_osd
*osd
, *nosd
;
1168 dout("__remove_old_osds %p\n", osdc
);
1169 mutex_lock(&osdc
->request_mutex
);
1170 list_for_each_entry_safe(osd
, nosd
, &osdc
->osd_lru
, o_osd_lru
) {
1171 if (time_before(jiffies
, osd
->lru_ttl
))
1173 remove_osd(osdc
, osd
);
1175 mutex_unlock(&osdc
->request_mutex
);
1181 static int __reset_osd(struct ceph_osd_client
*osdc
, struct ceph_osd
*osd
)
1183 struct ceph_entity_addr
*peer_addr
;
1185 dout("__reset_osd %p osd%d\n", osd
, osd
->o_osd
);
1186 if (list_empty(&osd
->o_requests
) &&
1187 list_empty(&osd
->o_linger_requests
)) {
1188 remove_osd(osdc
, osd
);
1192 peer_addr
= &osdc
->osdmap
->osd_addr
[osd
->o_osd
];
1193 if (!memcmp(peer_addr
, &osd
->o_con
.peer_addr
, sizeof (*peer_addr
)) &&
1194 !ceph_con_opened(&osd
->o_con
)) {
1195 struct ceph_osd_request
*req
;
1197 dout("osd addr hasn't changed and connection never opened, "
1198 "letting msgr retry\n");
1199 /* touch each r_stamp for handle_timeout()'s benfit */
1200 list_for_each_entry(req
, &osd
->o_requests
, r_osd_item
)
1201 req
->r_stamp
= jiffies
;
1206 ceph_con_close(&osd
->o_con
);
1207 ceph_con_open(&osd
->o_con
, CEPH_ENTITY_TYPE_OSD
, osd
->o_osd
, peer_addr
);
1208 osd
->o_incarnation
++;
1213 static void __insert_osd(struct ceph_osd_client
*osdc
, struct ceph_osd
*new)
1215 struct rb_node
**p
= &osdc
->osds
.rb_node
;
1216 struct rb_node
*parent
= NULL
;
1217 struct ceph_osd
*osd
= NULL
;
1219 dout("__insert_osd %p osd%d\n", new, new->o_osd
);
1222 osd
= rb_entry(parent
, struct ceph_osd
, o_node
);
1223 if (new->o_osd
< osd
->o_osd
)
1225 else if (new->o_osd
> osd
->o_osd
)
1226 p
= &(*p
)->rb_right
;
1231 rb_link_node(&new->o_node
, parent
, p
);
1232 rb_insert_color(&new->o_node
, &osdc
->osds
);
1235 static struct ceph_osd
*__lookup_osd(struct ceph_osd_client
*osdc
, int o
)
1237 struct ceph_osd
*osd
;
1238 struct rb_node
*n
= osdc
->osds
.rb_node
;
1241 osd
= rb_entry(n
, struct ceph_osd
, o_node
);
1244 else if (o
> osd
->o_osd
)
1252 static void __schedule_osd_timeout(struct ceph_osd_client
*osdc
)
1254 schedule_delayed_work(&osdc
->timeout_work
,
1255 osdc
->client
->options
->osd_keepalive_timeout
);
1258 static void __cancel_osd_timeout(struct ceph_osd_client
*osdc
)
1260 cancel_delayed_work(&osdc
->timeout_work
);
1264 * Register request, assign tid. If this is the first request, set up
1265 * the timeout event.
1267 static void __register_request(struct ceph_osd_client
*osdc
,
1268 struct ceph_osd_request
*req
)
1270 req
->r_tid
= ++osdc
->last_tid
;
1271 req
->r_request
->hdr
.tid
= cpu_to_le64(req
->r_tid
);
1272 dout("__register_request %p tid %lld\n", req
, req
->r_tid
);
1273 __insert_request(osdc
, req
);
1274 ceph_osdc_get_request(req
);
1275 osdc
->num_requests
++;
1276 if (osdc
->num_requests
== 1) {
1277 dout(" first request, scheduling timeout\n");
1278 __schedule_osd_timeout(osdc
);
1283 * called under osdc->request_mutex
1285 static void __unregister_request(struct ceph_osd_client
*osdc
,
1286 struct ceph_osd_request
*req
)
1288 if (RB_EMPTY_NODE(&req
->r_node
)) {
1289 dout("__unregister_request %p tid %lld not registered\n",
1294 dout("__unregister_request %p tid %lld\n", req
, req
->r_tid
);
1295 rb_erase(&req
->r_node
, &osdc
->requests
);
1296 RB_CLEAR_NODE(&req
->r_node
);
1297 osdc
->num_requests
--;
1300 /* make sure the original request isn't in flight. */
1301 ceph_msg_revoke(req
->r_request
);
1303 list_del_init(&req
->r_osd_item
);
1304 maybe_move_osd_to_lru(osdc
, req
->r_osd
);
1305 if (list_empty(&req
->r_linger_osd_item
))
1309 list_del_init(&req
->r_req_lru_item
);
1310 ceph_osdc_put_request(req
);
1312 if (osdc
->num_requests
== 0) {
1313 dout(" no requests, canceling timeout\n");
1314 __cancel_osd_timeout(osdc
);
1319 * Cancel a previously queued request message
1321 static void __cancel_request(struct ceph_osd_request
*req
)
1323 if (req
->r_sent
&& req
->r_osd
) {
1324 ceph_msg_revoke(req
->r_request
);
1329 static void __register_linger_request(struct ceph_osd_client
*osdc
,
1330 struct ceph_osd_request
*req
)
1332 dout("%s %p tid %llu\n", __func__
, req
, req
->r_tid
);
1333 WARN_ON(!req
->r_linger
);
1335 ceph_osdc_get_request(req
);
1336 list_add_tail(&req
->r_linger_item
, &osdc
->req_linger
);
1338 list_add_tail(&req
->r_linger_osd_item
,
1339 &req
->r_osd
->o_linger_requests
);
1342 static void __unregister_linger_request(struct ceph_osd_client
*osdc
,
1343 struct ceph_osd_request
*req
)
1345 WARN_ON(!req
->r_linger
);
1347 if (list_empty(&req
->r_linger_item
)) {
1348 dout("%s %p tid %llu not registered\n", __func__
, req
,
1353 dout("%s %p tid %llu\n", __func__
, req
, req
->r_tid
);
1354 list_del_init(&req
->r_linger_item
);
1357 list_del_init(&req
->r_linger_osd_item
);
1358 maybe_move_osd_to_lru(osdc
, req
->r_osd
);
1359 if (list_empty(&req
->r_osd_item
))
1362 ceph_osdc_put_request(req
);
1365 void ceph_osdc_set_request_linger(struct ceph_osd_client
*osdc
,
1366 struct ceph_osd_request
*req
)
1368 if (!req
->r_linger
) {
1369 dout("set_request_linger %p\n", req
);
1373 EXPORT_SYMBOL(ceph_osdc_set_request_linger
);
1376 * Returns whether a request should be blocked from being sent
1377 * based on the current osdmap and osd_client settings.
1379 * Caller should hold map_sem for read.
1381 static bool __req_should_be_paused(struct ceph_osd_client
*osdc
,
1382 struct ceph_osd_request
*req
)
1384 bool pauserd
= ceph_osdmap_flag(osdc
->osdmap
, CEPH_OSDMAP_PAUSERD
);
1385 bool pausewr
= ceph_osdmap_flag(osdc
->osdmap
, CEPH_OSDMAP_PAUSEWR
) ||
1386 ceph_osdmap_flag(osdc
->osdmap
, CEPH_OSDMAP_FULL
);
1387 return (req
->r_flags
& CEPH_OSD_FLAG_READ
&& pauserd
) ||
1388 (req
->r_flags
& CEPH_OSD_FLAG_WRITE
&& pausewr
);
1392 * Calculate mapping of a request to a PG. Takes tiering into account.
1394 static int __calc_request_pg(struct ceph_osdmap
*osdmap
,
1395 struct ceph_osd_request
*req
,
1396 struct ceph_pg
*pg_out
)
1398 bool need_check_tiering
;
1400 need_check_tiering
= false;
1401 if (req
->r_target_oloc
.pool
== -1) {
1402 req
->r_target_oloc
= req
->r_base_oloc
; /* struct */
1403 need_check_tiering
= true;
1405 if (req
->r_target_oid
.name_len
== 0) {
1406 ceph_oid_copy(&req
->r_target_oid
, &req
->r_base_oid
);
1407 need_check_tiering
= true;
1410 if (need_check_tiering
&&
1411 (req
->r_flags
& CEPH_OSD_FLAG_IGNORE_OVERLAY
) == 0) {
1412 struct ceph_pg_pool_info
*pi
;
1414 pi
= ceph_pg_pool_by_id(osdmap
, req
->r_target_oloc
.pool
);
1416 if ((req
->r_flags
& CEPH_OSD_FLAG_READ
) &&
1418 req
->r_target_oloc
.pool
= pi
->read_tier
;
1419 if ((req
->r_flags
& CEPH_OSD_FLAG_WRITE
) &&
1420 pi
->write_tier
>= 0)
1421 req
->r_target_oloc
.pool
= pi
->write_tier
;
1423 /* !pi is caught in ceph_oloc_oid_to_pg() */
1426 return ceph_oloc_oid_to_pg(osdmap
, &req
->r_target_oloc
,
1427 &req
->r_target_oid
, pg_out
);
1430 static void __enqueue_request(struct ceph_osd_request
*req
)
1432 struct ceph_osd_client
*osdc
= req
->r_osdc
;
1434 dout("%s %p tid %llu to osd%d\n", __func__
, req
, req
->r_tid
,
1435 req
->r_osd
? req
->r_osd
->o_osd
: -1);
1438 __remove_osd_from_lru(req
->r_osd
);
1439 list_add_tail(&req
->r_osd_item
, &req
->r_osd
->o_requests
);
1440 list_move_tail(&req
->r_req_lru_item
, &osdc
->req_unsent
);
1442 list_move_tail(&req
->r_req_lru_item
, &osdc
->req_notarget
);
1447 * Pick an osd (the first 'up' osd in the pg), allocate the osd struct
1448 * (as needed), and set the request r_osd appropriately. If there is
1449 * no up osd, set r_osd to NULL. Move the request to the appropriate list
1450 * (unsent, homeless) or leave on in-flight lru.
1452 * Return 0 if unchanged, 1 if changed, or negative on error.
1454 * Caller should hold map_sem for read and request_mutex.
1456 static int __map_request(struct ceph_osd_client
*osdc
,
1457 struct ceph_osd_request
*req
, int force_resend
)
1459 struct ceph_pg pgid
;
1460 int acting
[CEPH_PG_MAX_SIZE
];
1465 dout("map_request %p tid %lld\n", req
, req
->r_tid
);
1467 err
= __calc_request_pg(osdc
->osdmap
, req
, &pgid
);
1469 list_move(&req
->r_req_lru_item
, &osdc
->req_notarget
);
1474 num
= ceph_calc_pg_acting(osdc
->osdmap
, pgid
, acting
, &o
);
1478 was_paused
= req
->r_paused
;
1479 req
->r_paused
= __req_should_be_paused(osdc
, req
);
1480 if (was_paused
&& !req
->r_paused
)
1483 if ((!force_resend
&&
1484 req
->r_osd
&& req
->r_osd
->o_osd
== o
&&
1485 req
->r_sent
>= req
->r_osd
->o_incarnation
&&
1486 req
->r_num_pg_osds
== num
&&
1487 memcmp(req
->r_pg_osds
, acting
, sizeof(acting
[0])*num
) == 0) ||
1488 (req
->r_osd
== NULL
&& o
== -1) ||
1490 return 0; /* no change */
1492 dout("map_request tid %llu pgid %lld.%x osd%d (was osd%d)\n",
1493 req
->r_tid
, pgid
.pool
, pgid
.seed
, o
,
1494 req
->r_osd
? req
->r_osd
->o_osd
: -1);
1496 /* record full pg acting set */
1497 memcpy(req
->r_pg_osds
, acting
, sizeof(acting
[0]) * num
);
1498 req
->r_num_pg_osds
= num
;
1501 __cancel_request(req
);
1502 list_del_init(&req
->r_osd_item
);
1503 list_del_init(&req
->r_linger_osd_item
);
1507 req
->r_osd
= __lookup_osd(osdc
, o
);
1508 if (!req
->r_osd
&& o
>= 0) {
1510 req
->r_osd
= create_osd(osdc
, o
);
1512 list_move(&req
->r_req_lru_item
, &osdc
->req_notarget
);
1516 dout("map_request osd %p is osd%d\n", req
->r_osd
, o
);
1517 __insert_osd(osdc
, req
->r_osd
);
1519 ceph_con_open(&req
->r_osd
->o_con
,
1520 CEPH_ENTITY_TYPE_OSD
, o
,
1521 &osdc
->osdmap
->osd_addr
[o
]);
1524 __enqueue_request(req
);
1525 err
= 1; /* osd or pg changed */
1532 * caller should hold map_sem (for read) and request_mutex
1534 static void __send_request(struct ceph_osd_client
*osdc
,
1535 struct ceph_osd_request
*req
)
1539 dout("send_request %p tid %llu to osd%d flags %d pg %lld.%x\n",
1540 req
, req
->r_tid
, req
->r_osd
->o_osd
, req
->r_flags
,
1541 (unsigned long long)req
->r_pgid
.pool
, req
->r_pgid
.seed
);
1543 /* fill in message content that changes each time we send it */
1544 put_unaligned_le32(osdc
->osdmap
->epoch
, req
->r_request_osdmap_epoch
);
1545 put_unaligned_le32(req
->r_flags
, req
->r_request_flags
);
1546 put_unaligned_le64(req
->r_target_oloc
.pool
, req
->r_request_pool
);
1547 p
= req
->r_request_pgid
;
1548 ceph_encode_64(&p
, req
->r_pgid
.pool
);
1549 ceph_encode_32(&p
, req
->r_pgid
.seed
);
1550 put_unaligned_le64(1, req
->r_request_attempts
); /* FIXME */
1551 memcpy(req
->r_request_reassert_version
, &req
->r_reassert_version
,
1552 sizeof(req
->r_reassert_version
));
1554 req
->r_stamp
= jiffies
;
1555 list_move_tail(&req
->r_req_lru_item
, &osdc
->req_lru
);
1557 ceph_msg_get(req
->r_request
); /* send consumes a ref */
1559 req
->r_sent
= req
->r_osd
->o_incarnation
;
1561 ceph_con_send(&req
->r_osd
->o_con
, req
->r_request
);
1565 * Send any requests in the queue (req_unsent).
1567 static void __send_queued(struct ceph_osd_client
*osdc
)
1569 struct ceph_osd_request
*req
, *tmp
;
1571 dout("__send_queued\n");
1572 list_for_each_entry_safe(req
, tmp
, &osdc
->req_unsent
, r_req_lru_item
)
1573 __send_request(osdc
, req
);
1577 * Caller should hold map_sem for read and request_mutex.
1579 static int __ceph_osdc_start_request(struct ceph_osd_client
*osdc
,
1580 struct ceph_osd_request
*req
,
1585 __register_request(osdc
, req
);
1587 req
->r_got_reply
= 0;
1588 rc
= __map_request(osdc
, req
, 0);
1591 dout("osdc_start_request failed map, "
1592 " will retry %lld\n", req
->r_tid
);
1595 __unregister_request(osdc
, req
);
1600 if (req
->r_osd
== NULL
) {
1601 dout("send_request %p no up osds in pg\n", req
);
1602 ceph_monc_request_next_osdmap(&osdc
->client
->monc
);
1604 __send_queued(osdc
);
1611 * Timeout callback, called every N seconds when 1 or more osd
1612 * requests has been active for more than N seconds. When this
1613 * happens, we ping all OSDs with requests who have timed out to
1614 * ensure any communications channel reset is detected. Reset the
1615 * request timeouts another N seconds in the future as we go.
1616 * Reschedule the timeout event another N seconds in future (unless
1617 * there are no open requests).
1619 static void handle_timeout(struct work_struct
*work
)
1621 struct ceph_osd_client
*osdc
=
1622 container_of(work
, struct ceph_osd_client
, timeout_work
.work
);
1623 struct ceph_options
*opts
= osdc
->client
->options
;
1624 struct ceph_osd_request
*req
;
1625 struct ceph_osd
*osd
;
1626 struct list_head slow_osds
;
1628 down_read(&osdc
->map_sem
);
1630 ceph_monc_request_next_osdmap(&osdc
->client
->monc
);
1632 mutex_lock(&osdc
->request_mutex
);
1635 * ping osds that are a bit slow. this ensures that if there
1636 * is a break in the TCP connection we will notice, and reopen
1637 * a connection with that osd (from the fault callback).
1639 INIT_LIST_HEAD(&slow_osds
);
1640 list_for_each_entry(req
, &osdc
->req_lru
, r_req_lru_item
) {
1641 if (time_before(jiffies
,
1642 req
->r_stamp
+ opts
->osd_keepalive_timeout
))
1647 dout(" tid %llu is slow, will send keepalive on osd%d\n",
1648 req
->r_tid
, osd
->o_osd
);
1649 list_move_tail(&osd
->o_keepalive_item
, &slow_osds
);
1651 while (!list_empty(&slow_osds
)) {
1652 osd
= list_entry(slow_osds
.next
, struct ceph_osd
,
1654 list_del_init(&osd
->o_keepalive_item
);
1655 ceph_con_keepalive(&osd
->o_con
);
1658 __schedule_osd_timeout(osdc
);
1659 __send_queued(osdc
);
1660 mutex_unlock(&osdc
->request_mutex
);
1661 up_read(&osdc
->map_sem
);
1664 static void handle_osds_timeout(struct work_struct
*work
)
1666 struct ceph_osd_client
*osdc
=
1667 container_of(work
, struct ceph_osd_client
,
1668 osds_timeout_work
.work
);
1669 unsigned long delay
= osdc
->client
->options
->osd_idle_ttl
/ 4;
1671 dout("osds timeout\n");
1672 down_read(&osdc
->map_sem
);
1673 remove_old_osds(osdc
);
1674 up_read(&osdc
->map_sem
);
1676 schedule_delayed_work(&osdc
->osds_timeout_work
,
1677 round_jiffies_relative(delay
));
1680 static int ceph_oloc_decode(void **p
, void *end
,
1681 struct ceph_object_locator
*oloc
)
1683 u8 struct_v
, struct_cv
;
1688 ceph_decode_need(p
, end
, 1 + 1 + 4, e_inval
);
1689 struct_v
= ceph_decode_8(p
);
1690 struct_cv
= ceph_decode_8(p
);
1692 pr_warn("got v %d < 3 cv %d of ceph_object_locator\n",
1693 struct_v
, struct_cv
);
1696 if (struct_cv
> 6) {
1697 pr_warn("got v %d cv %d > 6 of ceph_object_locator\n",
1698 struct_v
, struct_cv
);
1701 len
= ceph_decode_32(p
);
1702 ceph_decode_need(p
, end
, len
, e_inval
);
1703 struct_end
= *p
+ len
;
1705 oloc
->pool
= ceph_decode_64(p
);
1706 *p
+= 4; /* skip preferred */
1708 len
= ceph_decode_32(p
);
1710 pr_warn("ceph_object_locator::key is set\n");
1714 if (struct_v
>= 5) {
1715 len
= ceph_decode_32(p
);
1717 pr_warn("ceph_object_locator::nspace is set\n");
1722 if (struct_v
>= 6) {
1723 s64 hash
= ceph_decode_64(p
);
1725 pr_warn("ceph_object_locator::hash is set\n");
1740 static int ceph_redirect_decode(void **p
, void *end
,
1741 struct ceph_request_redirect
*redir
)
1743 u8 struct_v
, struct_cv
;
1748 ceph_decode_need(p
, end
, 1 + 1 + 4, e_inval
);
1749 struct_v
= ceph_decode_8(p
);
1750 struct_cv
= ceph_decode_8(p
);
1751 if (struct_cv
> 1) {
1752 pr_warn("got v %d cv %d > 1 of ceph_request_redirect\n",
1753 struct_v
, struct_cv
);
1756 len
= ceph_decode_32(p
);
1757 ceph_decode_need(p
, end
, len
, e_inval
);
1758 struct_end
= *p
+ len
;
1760 ret
= ceph_oloc_decode(p
, end
, &redir
->oloc
);
1764 len
= ceph_decode_32(p
);
1766 pr_warn("ceph_request_redirect::object_name is set\n");
1770 len
= ceph_decode_32(p
);
1771 *p
+= len
; /* skip osd_instructions */
1783 static void complete_request(struct ceph_osd_request
*req
)
1785 complete_all(&req
->r_safe_completion
); /* fsync waiter */
1789 * handle osd op reply. either call the callback if it is specified,
1790 * or do the completion to wake up the waiting thread.
1792 static void handle_reply(struct ceph_osd_client
*osdc
, struct ceph_msg
*msg
)
1795 struct ceph_osd_request
*req
;
1796 struct ceph_request_redirect redir
;
1799 unsigned int numops
;
1800 int payload_len
, flags
;
1806 u64 reassert_version
;
1808 int already_completed
;
1813 tid
= le64_to_cpu(msg
->hdr
.tid
);
1814 dout("handle_reply %p tid %llu\n", msg
, tid
);
1816 p
= msg
->front
.iov_base
;
1817 end
= p
+ msg
->front
.iov_len
;
1819 ceph_decode_need(&p
, end
, 4, bad
);
1820 object_len
= ceph_decode_32(&p
);
1821 ceph_decode_need(&p
, end
, object_len
, bad
);
1824 err
= ceph_decode_pgid(&p
, end
, &pg
);
1828 ceph_decode_need(&p
, end
, 8 + 4 + 4 + 8 + 4, bad
);
1829 flags
= ceph_decode_64(&p
);
1830 result
= ceph_decode_32(&p
);
1831 reassert_epoch
= ceph_decode_32(&p
);
1832 reassert_version
= ceph_decode_64(&p
);
1833 osdmap_epoch
= ceph_decode_32(&p
);
1836 down_read(&osdc
->map_sem
);
1837 mutex_lock(&osdc
->request_mutex
);
1838 req
= __lookup_request(osdc
, tid
);
1840 dout("handle_reply tid %llu dne\n", tid
);
1843 ceph_osdc_get_request(req
);
1845 dout("handle_reply %p tid %llu req %p result %d\n", msg
, tid
,
1848 ceph_decode_need(&p
, end
, 4, bad_put
);
1849 numops
= ceph_decode_32(&p
);
1850 if (numops
> CEPH_OSD_MAX_OPS
)
1852 if (numops
!= req
->r_num_ops
)
1855 ceph_decode_need(&p
, end
, numops
* sizeof(struct ceph_osd_op
), bad_put
);
1856 for (i
= 0; i
< numops
; i
++) {
1857 struct ceph_osd_op
*op
= p
;
1860 len
= le32_to_cpu(op
->payload_len
);
1861 req
->r_ops
[i
].outdata_len
= len
;
1862 dout(" op %d has %d bytes\n", i
, len
);
1866 bytes
= le32_to_cpu(msg
->hdr
.data_len
);
1867 if (payload_len
!= bytes
) {
1868 pr_warn("sum of op payload lens %d != data_len %d\n",
1869 payload_len
, bytes
);
1873 ceph_decode_need(&p
, end
, 4 + numops
* 4, bad_put
);
1874 retry_attempt
= ceph_decode_32(&p
);
1875 for (i
= 0; i
< numops
; i
++)
1876 req
->r_ops
[i
].rval
= ceph_decode_32(&p
);
1878 if (le16_to_cpu(msg
->hdr
.version
) >= 6) {
1879 p
+= 8 + 4; /* skip replay_version */
1880 p
+= 8; /* skip user_version */
1882 if (le16_to_cpu(msg
->hdr
.version
) >= 7)
1883 ceph_decode_8_safe(&p
, end
, decode_redir
, bad_put
);
1891 err
= ceph_redirect_decode(&p
, end
, &redir
);
1895 redir
.oloc
.pool
= -1;
1898 if (redir
.oloc
.pool
!= -1) {
1899 dout("redirect pool %lld\n", redir
.oloc
.pool
);
1901 __unregister_request(osdc
, req
);
1903 req
->r_target_oloc
= redir
.oloc
; /* struct */
1906 * Start redirect requests with nofail=true. If
1907 * mapping fails, request will end up on the notarget
1908 * list, waiting for the new osdmap (which can take
1909 * a while), even though the original request mapped
1910 * successfully. In the future we might want to follow
1911 * original request's nofail setting here.
1913 err
= __ceph_osdc_start_request(osdc
, req
, true);
1919 already_completed
= req
->r_got_reply
;
1920 if (!req
->r_got_reply
) {
1921 req
->r_result
= result
;
1922 dout("handle_reply result %d bytes %d\n", req
->r_result
,
1924 if (req
->r_result
== 0)
1925 req
->r_result
= bytes
;
1927 /* in case this is a write and we need to replay, */
1928 req
->r_reassert_version
.epoch
= cpu_to_le32(reassert_epoch
);
1929 req
->r_reassert_version
.version
= cpu_to_le64(reassert_version
);
1931 req
->r_got_reply
= 1;
1932 } else if ((flags
& CEPH_OSD_FLAG_ONDISK
) == 0) {
1933 dout("handle_reply tid %llu dup ack\n", tid
);
1937 dout("handle_reply tid %llu flags %d\n", tid
, flags
);
1939 if (req
->r_linger
&& (flags
& CEPH_OSD_FLAG_ONDISK
))
1940 __register_linger_request(osdc
, req
);
1942 /* either this is a read, or we got the safe response */
1944 (flags
& CEPH_OSD_FLAG_ONDISK
) ||
1945 ((flags
& CEPH_OSD_FLAG_WRITE
) == 0))
1946 __unregister_request(osdc
, req
);
1948 mutex_unlock(&osdc
->request_mutex
);
1949 up_read(&osdc
->map_sem
);
1951 if (!already_completed
) {
1952 if (req
->r_unsafe_callback
&&
1953 result
>= 0 && !(flags
& CEPH_OSD_FLAG_ONDISK
))
1954 req
->r_unsafe_callback(req
, true);
1955 if (req
->r_callback
)
1956 req
->r_callback(req
, msg
);
1958 complete_all(&req
->r_completion
);
1961 if (flags
& CEPH_OSD_FLAG_ONDISK
) {
1962 if (req
->r_unsafe_callback
&& already_completed
)
1963 req
->r_unsafe_callback(req
, false);
1964 complete_request(req
);
1968 dout("req=%p req->r_linger=%d\n", req
, req
->r_linger
);
1969 ceph_osdc_put_request(req
);
1972 mutex_unlock(&osdc
->request_mutex
);
1973 up_read(&osdc
->map_sem
);
1977 req
->r_result
= -EIO
;
1978 __unregister_request(osdc
, req
);
1979 if (req
->r_callback
)
1980 req
->r_callback(req
, msg
);
1982 complete_all(&req
->r_completion
);
1983 complete_request(req
);
1984 ceph_osdc_put_request(req
);
1986 mutex_unlock(&osdc
->request_mutex
);
1987 up_read(&osdc
->map_sem
);
1989 pr_err("corrupt osd_op_reply got %d %d\n",
1990 (int)msg
->front
.iov_len
, le32_to_cpu(msg
->hdr
.front_len
));
1994 static void reset_changed_osds(struct ceph_osd_client
*osdc
)
1996 struct rb_node
*p
, *n
;
1998 dout("%s %p\n", __func__
, osdc
);
1999 for (p
= rb_first(&osdc
->osds
); p
; p
= n
) {
2000 struct ceph_osd
*osd
= rb_entry(p
, struct ceph_osd
, o_node
);
2003 if (!ceph_osd_is_up(osdc
->osdmap
, osd
->o_osd
) ||
2004 memcmp(&osd
->o_con
.peer_addr
,
2005 ceph_osd_addr(osdc
->osdmap
,
2007 sizeof(struct ceph_entity_addr
)) != 0)
2008 __reset_osd(osdc
, osd
);
2013 * Requeue requests whose mapping to an OSD has changed. If requests map to
2014 * no osd, request a new map.
2016 * Caller should hold map_sem for read.
2018 static void kick_requests(struct ceph_osd_client
*osdc
, bool force_resend
,
2019 bool force_resend_writes
)
2021 struct ceph_osd_request
*req
, *nreq
;
2025 bool force_resend_req
;
2027 dout("kick_requests %s %s\n", force_resend
? " (force resend)" : "",
2028 force_resend_writes
? " (force resend writes)" : "");
2029 mutex_lock(&osdc
->request_mutex
);
2030 for (p
= rb_first(&osdc
->requests
); p
; ) {
2031 req
= rb_entry(p
, struct ceph_osd_request
, r_node
);
2035 * For linger requests that have not yet been
2036 * registered, move them to the linger list; they'll
2037 * be sent to the osd in the loop below. Unregister
2038 * the request before re-registering it as a linger
2039 * request to ensure the __map_request() below
2040 * will decide it needs to be sent.
2042 if (req
->r_linger
&& list_empty(&req
->r_linger_item
)) {
2043 dout("%p tid %llu restart on osd%d\n",
2045 req
->r_osd
? req
->r_osd
->o_osd
: -1);
2046 ceph_osdc_get_request(req
);
2047 __unregister_request(osdc
, req
);
2048 __register_linger_request(osdc
, req
);
2049 ceph_osdc_put_request(req
);
2053 force_resend_req
= force_resend
||
2054 (force_resend_writes
&&
2055 req
->r_flags
& CEPH_OSD_FLAG_WRITE
);
2056 err
= __map_request(osdc
, req
, force_resend_req
);
2058 continue; /* error */
2059 if (req
->r_osd
== NULL
) {
2060 dout("%p tid %llu maps to no osd\n", req
, req
->r_tid
);
2061 needmap
++; /* request a newer map */
2062 } else if (err
> 0) {
2063 if (!req
->r_linger
) {
2064 dout("%p tid %llu requeued on osd%d\n", req
,
2066 req
->r_osd
? req
->r_osd
->o_osd
: -1);
2067 req
->r_flags
|= CEPH_OSD_FLAG_RETRY
;
2072 list_for_each_entry_safe(req
, nreq
, &osdc
->req_linger
,
2074 dout("linger req=%p req->r_osd=%p\n", req
, req
->r_osd
);
2076 err
= __map_request(osdc
, req
,
2077 force_resend
|| force_resend_writes
);
2078 dout("__map_request returned %d\n", err
);
2080 continue; /* hrm! */
2081 if (req
->r_osd
== NULL
|| err
> 0) {
2082 if (req
->r_osd
== NULL
) {
2083 dout("lingering %p tid %llu maps to no osd\n",
2086 * A homeless lingering request makes
2087 * no sense, as it's job is to keep
2088 * a particular OSD connection open.
2089 * Request a newer map and kick the
2090 * request, knowing that it won't be
2091 * resent until we actually get a map
2092 * that can tell us where to send it.
2097 dout("kicking lingering %p tid %llu osd%d\n", req
,
2098 req
->r_tid
, req
->r_osd
? req
->r_osd
->o_osd
: -1);
2099 __register_request(osdc
, req
);
2100 __unregister_linger_request(osdc
, req
);
2103 reset_changed_osds(osdc
);
2104 mutex_unlock(&osdc
->request_mutex
);
2107 dout("%d requests for down osds, need new map\n", needmap
);
2108 ceph_monc_request_next_osdmap(&osdc
->client
->monc
);
2114 * Process updated osd map.
2116 * The message contains any number of incremental and full maps, normally
2117 * indicating some sort of topology change in the cluster. Kick requests
2118 * off to different OSDs as needed.
2120 void ceph_osdc_handle_map(struct ceph_osd_client
*osdc
, struct ceph_msg
*msg
)
2122 void *p
, *end
, *next
;
2123 u32 nr_maps
, maplen
;
2125 struct ceph_osdmap
*newmap
= NULL
, *oldmap
;
2127 struct ceph_fsid fsid
;
2130 dout("handle_map have %u\n", osdc
->osdmap
? osdc
->osdmap
->epoch
: 0);
2131 p
= msg
->front
.iov_base
;
2132 end
= p
+ msg
->front
.iov_len
;
2135 ceph_decode_need(&p
, end
, sizeof(fsid
), bad
);
2136 ceph_decode_copy(&p
, &fsid
, sizeof(fsid
));
2137 if (ceph_check_fsid(osdc
->client
, &fsid
) < 0)
2140 down_write(&osdc
->map_sem
);
2142 was_full
= ceph_osdmap_flag(osdc
->osdmap
, CEPH_OSDMAP_FULL
);
2144 /* incremental maps */
2145 ceph_decode_32_safe(&p
, end
, nr_maps
, bad
);
2146 dout(" %d inc maps\n", nr_maps
);
2147 while (nr_maps
> 0) {
2148 ceph_decode_need(&p
, end
, 2*sizeof(u32
), bad
);
2149 epoch
= ceph_decode_32(&p
);
2150 maplen
= ceph_decode_32(&p
);
2151 ceph_decode_need(&p
, end
, maplen
, bad
);
2153 if (osdc
->osdmap
&& osdc
->osdmap
->epoch
+1 == epoch
) {
2154 dout("applying incremental map %u len %d\n",
2156 newmap
= osdmap_apply_incremental(&p
, next
,
2158 &osdc
->client
->msgr
);
2159 if (IS_ERR(newmap
)) {
2160 err
= PTR_ERR(newmap
);
2164 if (newmap
!= osdc
->osdmap
) {
2165 ceph_osdmap_destroy(osdc
->osdmap
);
2166 osdc
->osdmap
= newmap
;
2168 was_full
= was_full
||
2169 ceph_osdmap_flag(osdc
->osdmap
,
2171 kick_requests(osdc
, 0, was_full
);
2173 dout("ignoring incremental map %u len %d\n",
2183 ceph_decode_32_safe(&p
, end
, nr_maps
, bad
);
2184 dout(" %d full maps\n", nr_maps
);
2186 ceph_decode_need(&p
, end
, 2*sizeof(u32
), bad
);
2187 epoch
= ceph_decode_32(&p
);
2188 maplen
= ceph_decode_32(&p
);
2189 ceph_decode_need(&p
, end
, maplen
, bad
);
2191 dout("skipping non-latest full map %u len %d\n",
2193 } else if (osdc
->osdmap
&& osdc
->osdmap
->epoch
>= epoch
) {
2194 dout("skipping full map %u len %d, "
2195 "older than our %u\n", epoch
, maplen
,
2196 osdc
->osdmap
->epoch
);
2198 int skipped_map
= 0;
2200 dout("taking full map %u len %d\n", epoch
, maplen
);
2201 newmap
= ceph_osdmap_decode(&p
, p
+maplen
);
2202 if (IS_ERR(newmap
)) {
2203 err
= PTR_ERR(newmap
);
2207 oldmap
= osdc
->osdmap
;
2208 osdc
->osdmap
= newmap
;
2210 if (oldmap
->epoch
+ 1 < newmap
->epoch
)
2212 ceph_osdmap_destroy(oldmap
);
2214 was_full
= was_full
||
2215 ceph_osdmap_flag(osdc
->osdmap
,
2217 kick_requests(osdc
, skipped_map
, was_full
);
2226 downgrade_write(&osdc
->map_sem
);
2227 ceph_monc_got_map(&osdc
->client
->monc
, CEPH_SUB_OSDMAP
,
2228 osdc
->osdmap
->epoch
);
2231 * subscribe to subsequent osdmap updates if full to ensure
2232 * we find out when we are no longer full and stop returning
2235 if (ceph_osdmap_flag(osdc
->osdmap
, CEPH_OSDMAP_FULL
) ||
2236 ceph_osdmap_flag(osdc
->osdmap
, CEPH_OSDMAP_PAUSERD
) ||
2237 ceph_osdmap_flag(osdc
->osdmap
, CEPH_OSDMAP_PAUSEWR
))
2238 ceph_monc_request_next_osdmap(&osdc
->client
->monc
);
2240 mutex_lock(&osdc
->request_mutex
);
2241 __send_queued(osdc
);
2242 mutex_unlock(&osdc
->request_mutex
);
2243 up_read(&osdc
->map_sem
);
2244 wake_up_all(&osdc
->client
->auth_wq
);
2248 pr_err("osdc handle_map corrupt msg\n");
2250 up_write(&osdc
->map_sem
);
2254 * watch/notify callback event infrastructure
2256 * These callbacks are used both for watch and notify operations.
2258 static void __release_event(struct kref
*kref
)
2260 struct ceph_osd_event
*event
=
2261 container_of(kref
, struct ceph_osd_event
, kref
);
2263 dout("__release_event %p\n", event
);
2267 static void get_event(struct ceph_osd_event
*event
)
2269 kref_get(&event
->kref
);
2272 void ceph_osdc_put_event(struct ceph_osd_event
*event
)
2274 kref_put(&event
->kref
, __release_event
);
2276 EXPORT_SYMBOL(ceph_osdc_put_event
);
2278 static void __insert_event(struct ceph_osd_client
*osdc
,
2279 struct ceph_osd_event
*new)
2281 struct rb_node
**p
= &osdc
->event_tree
.rb_node
;
2282 struct rb_node
*parent
= NULL
;
2283 struct ceph_osd_event
*event
= NULL
;
2287 event
= rb_entry(parent
, struct ceph_osd_event
, node
);
2288 if (new->cookie
< event
->cookie
)
2290 else if (new->cookie
> event
->cookie
)
2291 p
= &(*p
)->rb_right
;
2296 rb_link_node(&new->node
, parent
, p
);
2297 rb_insert_color(&new->node
, &osdc
->event_tree
);
2300 static struct ceph_osd_event
*__find_event(struct ceph_osd_client
*osdc
,
2303 struct rb_node
**p
= &osdc
->event_tree
.rb_node
;
2304 struct rb_node
*parent
= NULL
;
2305 struct ceph_osd_event
*event
= NULL
;
2309 event
= rb_entry(parent
, struct ceph_osd_event
, node
);
2310 if (cookie
< event
->cookie
)
2312 else if (cookie
> event
->cookie
)
2313 p
= &(*p
)->rb_right
;
2320 static void __remove_event(struct ceph_osd_event
*event
)
2322 struct ceph_osd_client
*osdc
= event
->osdc
;
2324 if (!RB_EMPTY_NODE(&event
->node
)) {
2325 dout("__remove_event removed %p\n", event
);
2326 rb_erase(&event
->node
, &osdc
->event_tree
);
2327 ceph_osdc_put_event(event
);
2329 dout("__remove_event didn't remove %p\n", event
);
2333 int ceph_osdc_create_event(struct ceph_osd_client
*osdc
,
2334 void (*event_cb
)(u64
, u64
, u8
, void *),
2335 void *data
, struct ceph_osd_event
**pevent
)
2337 struct ceph_osd_event
*event
;
2339 event
= kmalloc(sizeof(*event
), GFP_NOIO
);
2343 dout("create_event %p\n", event
);
2344 event
->cb
= event_cb
;
2345 event
->one_shot
= 0;
2348 INIT_LIST_HEAD(&event
->osd_node
);
2349 RB_CLEAR_NODE(&event
->node
);
2350 kref_init(&event
->kref
); /* one ref for us */
2351 kref_get(&event
->kref
); /* one ref for the caller */
2353 spin_lock(&osdc
->event_lock
);
2354 event
->cookie
= ++osdc
->event_count
;
2355 __insert_event(osdc
, event
);
2356 spin_unlock(&osdc
->event_lock
);
2361 EXPORT_SYMBOL(ceph_osdc_create_event
);
2363 void ceph_osdc_cancel_event(struct ceph_osd_event
*event
)
2365 struct ceph_osd_client
*osdc
= event
->osdc
;
2367 dout("cancel_event %p\n", event
);
2368 spin_lock(&osdc
->event_lock
);
2369 __remove_event(event
);
2370 spin_unlock(&osdc
->event_lock
);
2371 ceph_osdc_put_event(event
); /* caller's */
2373 EXPORT_SYMBOL(ceph_osdc_cancel_event
);
2376 static void do_event_work(struct work_struct
*work
)
2378 struct ceph_osd_event_work
*event_work
=
2379 container_of(work
, struct ceph_osd_event_work
, work
);
2380 struct ceph_osd_event
*event
= event_work
->event
;
2381 u64 ver
= event_work
->ver
;
2382 u64 notify_id
= event_work
->notify_id
;
2383 u8 opcode
= event_work
->opcode
;
2385 dout("do_event_work completing %p\n", event
);
2386 event
->cb(ver
, notify_id
, opcode
, event
->data
);
2387 dout("do_event_work completed %p\n", event
);
2388 ceph_osdc_put_event(event
);
2394 * Process osd watch notifications
2396 static void handle_watch_notify(struct ceph_osd_client
*osdc
,
2397 struct ceph_msg
*msg
)
2401 u64 cookie
, ver
, notify_id
;
2403 struct ceph_osd_event
*event
;
2404 struct ceph_osd_event_work
*event_work
;
2406 p
= msg
->front
.iov_base
;
2407 end
= p
+ msg
->front
.iov_len
;
2409 ceph_decode_8_safe(&p
, end
, proto_ver
, bad
);
2410 ceph_decode_8_safe(&p
, end
, opcode
, bad
);
2411 ceph_decode_64_safe(&p
, end
, cookie
, bad
);
2412 ceph_decode_64_safe(&p
, end
, ver
, bad
);
2413 ceph_decode_64_safe(&p
, end
, notify_id
, bad
);
2415 spin_lock(&osdc
->event_lock
);
2416 event
= __find_event(osdc
, cookie
);
2418 BUG_ON(event
->one_shot
);
2421 spin_unlock(&osdc
->event_lock
);
2422 dout("handle_watch_notify cookie %lld ver %lld event %p\n",
2423 cookie
, ver
, event
);
2425 event_work
= kmalloc(sizeof(*event_work
), GFP_NOIO
);
2427 pr_err("couldn't allocate event_work\n");
2428 ceph_osdc_put_event(event
);
2431 INIT_WORK(&event_work
->work
, do_event_work
);
2432 event_work
->event
= event
;
2433 event_work
->ver
= ver
;
2434 event_work
->notify_id
= notify_id
;
2435 event_work
->opcode
= opcode
;
2437 queue_work(osdc
->notify_wq
, &event_work
->work
);
2443 pr_err("osdc handle_watch_notify corrupt msg\n");
2447 * build new request AND message
2450 void ceph_osdc_build_request(struct ceph_osd_request
*req
, u64 off
,
2451 struct ceph_snap_context
*snapc
, u64 snap_id
,
2452 struct timespec
*mtime
)
2454 struct ceph_msg
*msg
= req
->r_request
;
2457 int flags
= req
->r_flags
;
2461 req
->r_snapid
= snap_id
;
2462 WARN_ON(snapc
!= req
->r_snapc
);
2464 /* encode request */
2465 msg
->hdr
.version
= cpu_to_le16(4);
2467 p
= msg
->front
.iov_base
;
2468 ceph_encode_32(&p
, 1); /* client_inc is always 1 */
2469 req
->r_request_osdmap_epoch
= p
;
2471 req
->r_request_flags
= p
;
2473 if (req
->r_flags
& CEPH_OSD_FLAG_WRITE
)
2474 ceph_encode_timespec(p
, mtime
);
2475 p
+= sizeof(struct ceph_timespec
);
2476 req
->r_request_reassert_version
= p
;
2477 p
+= sizeof(struct ceph_eversion
); /* will get filled in */
2480 ceph_encode_8(&p
, 4);
2481 ceph_encode_8(&p
, 4);
2482 ceph_encode_32(&p
, 8 + 4 + 4);
2483 req
->r_request_pool
= p
;
2485 ceph_encode_32(&p
, -1); /* preferred */
2486 ceph_encode_32(&p
, 0); /* key len */
2488 ceph_encode_8(&p
, 1);
2489 req
->r_request_pgid
= p
;
2491 ceph_encode_32(&p
, -1); /* preferred */
2494 ceph_encode_32(&p
, req
->r_base_oid
.name_len
);
2495 memcpy(p
, req
->r_base_oid
.name
, req
->r_base_oid
.name_len
);
2496 dout("oid '%.*s' len %d\n", req
->r_base_oid
.name_len
,
2497 req
->r_base_oid
.name
, req
->r_base_oid
.name_len
);
2498 p
+= req
->r_base_oid
.name_len
;
2500 /* ops--can imply data */
2501 ceph_encode_16(&p
, (u16
)req
->r_num_ops
);
2503 for (i
= 0; i
< req
->r_num_ops
; i
++) {
2504 data_len
+= osd_req_encode_op(req
, p
, i
);
2505 p
+= sizeof(struct ceph_osd_op
);
2509 ceph_encode_64(&p
, req
->r_snapid
);
2510 ceph_encode_64(&p
, req
->r_snapc
? req
->r_snapc
->seq
: 0);
2511 ceph_encode_32(&p
, req
->r_snapc
? req
->r_snapc
->num_snaps
: 0);
2513 for (i
= 0; i
< req
->r_snapc
->num_snaps
; i
++) {
2514 ceph_encode_64(&p
, req
->r_snapc
->snaps
[i
]);
2518 req
->r_request_attempts
= p
;
2522 if (flags
& CEPH_OSD_FLAG_WRITE
) {
2526 * The header "data_off" is a hint to the receiver
2527 * allowing it to align received data into its
2528 * buffers such that there's no need to re-copy
2529 * it before writing it to disk (direct I/O).
2531 data_off
= (u16
) (off
& 0xffff);
2532 req
->r_request
->hdr
.data_off
= cpu_to_le16(data_off
);
2534 req
->r_request
->hdr
.data_len
= cpu_to_le32(data_len
);
2536 BUG_ON(p
> msg
->front
.iov_base
+ msg
->front
.iov_len
);
2537 msg_size
= p
- msg
->front
.iov_base
;
2538 msg
->front
.iov_len
= msg_size
;
2539 msg
->hdr
.front_len
= cpu_to_le32(msg_size
);
2541 dout("build_request msg_size was %d\n", (int)msg_size
);
2543 EXPORT_SYMBOL(ceph_osdc_build_request
);
2546 * Register request, send initial attempt.
2548 int ceph_osdc_start_request(struct ceph_osd_client
*osdc
,
2549 struct ceph_osd_request
*req
,
2554 down_read(&osdc
->map_sem
);
2555 mutex_lock(&osdc
->request_mutex
);
2557 rc
= __ceph_osdc_start_request(osdc
, req
, nofail
);
2559 mutex_unlock(&osdc
->request_mutex
);
2560 up_read(&osdc
->map_sem
);
2564 EXPORT_SYMBOL(ceph_osdc_start_request
);
2567 * Unregister a registered request. The request is not completed (i.e.
2568 * no callbacks or wakeups) - higher layers are supposed to know what
2569 * they are canceling.
2571 void ceph_osdc_cancel_request(struct ceph_osd_request
*req
)
2573 struct ceph_osd_client
*osdc
= req
->r_osdc
;
2575 mutex_lock(&osdc
->request_mutex
);
2577 __unregister_linger_request(osdc
, req
);
2578 __unregister_request(osdc
, req
);
2579 mutex_unlock(&osdc
->request_mutex
);
2581 dout("%s %p tid %llu canceled\n", __func__
, req
, req
->r_tid
);
2583 EXPORT_SYMBOL(ceph_osdc_cancel_request
);
2586 * wait for a request to complete
2588 int ceph_osdc_wait_request(struct ceph_osd_client
*osdc
,
2589 struct ceph_osd_request
*req
)
2593 dout("%s %p tid %llu\n", __func__
, req
, req
->r_tid
);
2595 rc
= wait_for_completion_interruptible(&req
->r_completion
);
2597 dout("%s %p tid %llu interrupted\n", __func__
, req
, req
->r_tid
);
2598 ceph_osdc_cancel_request(req
);
2599 complete_request(req
);
2603 dout("%s %p tid %llu result %d\n", __func__
, req
, req
->r_tid
,
2605 return req
->r_result
;
2607 EXPORT_SYMBOL(ceph_osdc_wait_request
);
2610 * sync - wait for all in-flight requests to flush. avoid starvation.
2612 void ceph_osdc_sync(struct ceph_osd_client
*osdc
)
2614 struct ceph_osd_request
*req
;
2615 u64 last_tid
, next_tid
= 0;
2617 mutex_lock(&osdc
->request_mutex
);
2618 last_tid
= osdc
->last_tid
;
2620 req
= __lookup_request_ge(osdc
, next_tid
);
2623 if (req
->r_tid
> last_tid
)
2626 next_tid
= req
->r_tid
+ 1;
2627 if ((req
->r_flags
& CEPH_OSD_FLAG_WRITE
) == 0)
2630 ceph_osdc_get_request(req
);
2631 mutex_unlock(&osdc
->request_mutex
);
2632 dout("sync waiting on tid %llu (last is %llu)\n",
2633 req
->r_tid
, last_tid
);
2634 wait_for_completion(&req
->r_safe_completion
);
2635 mutex_lock(&osdc
->request_mutex
);
2636 ceph_osdc_put_request(req
);
2638 mutex_unlock(&osdc
->request_mutex
);
2639 dout("sync done (thru tid %llu)\n", last_tid
);
2641 EXPORT_SYMBOL(ceph_osdc_sync
);
2644 * Call all pending notify callbacks - for use after a watch is
2645 * unregistered, to make sure no more callbacks for it will be invoked
2647 void ceph_osdc_flush_notifies(struct ceph_osd_client
*osdc
)
2649 flush_workqueue(osdc
->notify_wq
);
2651 EXPORT_SYMBOL(ceph_osdc_flush_notifies
);
2657 int ceph_osdc_init(struct ceph_osd_client
*osdc
, struct ceph_client
*client
)
2662 osdc
->client
= client
;
2663 osdc
->osdmap
= NULL
;
2664 init_rwsem(&osdc
->map_sem
);
2665 init_completion(&osdc
->map_waiters
);
2666 osdc
->last_requested_map
= 0;
2667 mutex_init(&osdc
->request_mutex
);
2669 osdc
->osds
= RB_ROOT
;
2670 INIT_LIST_HEAD(&osdc
->osd_lru
);
2671 osdc
->requests
= RB_ROOT
;
2672 INIT_LIST_HEAD(&osdc
->req_lru
);
2673 INIT_LIST_HEAD(&osdc
->req_unsent
);
2674 INIT_LIST_HEAD(&osdc
->req_notarget
);
2675 INIT_LIST_HEAD(&osdc
->req_linger
);
2676 osdc
->num_requests
= 0;
2677 INIT_DELAYED_WORK(&osdc
->timeout_work
, handle_timeout
);
2678 INIT_DELAYED_WORK(&osdc
->osds_timeout_work
, handle_osds_timeout
);
2679 spin_lock_init(&osdc
->event_lock
);
2680 osdc
->event_tree
= RB_ROOT
;
2681 osdc
->event_count
= 0;
2683 schedule_delayed_work(&osdc
->osds_timeout_work
,
2684 round_jiffies_relative(osdc
->client
->options
->osd_idle_ttl
));
2687 osdc
->req_mempool
= mempool_create_slab_pool(10,
2688 ceph_osd_request_cache
);
2689 if (!osdc
->req_mempool
)
2692 err
= ceph_msgpool_init(&osdc
->msgpool_op
, CEPH_MSG_OSD_OP
,
2693 OSD_OP_FRONT_LEN
, 10, true,
2697 err
= ceph_msgpool_init(&osdc
->msgpool_op_reply
, CEPH_MSG_OSD_OPREPLY
,
2698 OSD_OPREPLY_FRONT_LEN
, 10, true,
2704 osdc
->notify_wq
= create_singlethread_workqueue("ceph-watch-notify");
2705 if (!osdc
->notify_wq
)
2706 goto out_msgpool_reply
;
2711 ceph_msgpool_destroy(&osdc
->msgpool_op_reply
);
2713 ceph_msgpool_destroy(&osdc
->msgpool_op
);
2715 mempool_destroy(osdc
->req_mempool
);
2720 void ceph_osdc_stop(struct ceph_osd_client
*osdc
)
2722 flush_workqueue(osdc
->notify_wq
);
2723 destroy_workqueue(osdc
->notify_wq
);
2724 cancel_delayed_work_sync(&osdc
->timeout_work
);
2725 cancel_delayed_work_sync(&osdc
->osds_timeout_work
);
2727 ceph_osdmap_destroy(osdc
->osdmap
);
2728 osdc
->osdmap
= NULL
;
2730 remove_all_osds(osdc
);
2731 mempool_destroy(osdc
->req_mempool
);
2732 ceph_msgpool_destroy(&osdc
->msgpool_op
);
2733 ceph_msgpool_destroy(&osdc
->msgpool_op_reply
);
2737 * Read some contiguous pages. If we cross a stripe boundary, shorten
2738 * *plen. Return number of bytes read, or error.
2740 int ceph_osdc_readpages(struct ceph_osd_client
*osdc
,
2741 struct ceph_vino vino
, struct ceph_file_layout
*layout
,
2743 u32 truncate_seq
, u64 truncate_size
,
2744 struct page
**pages
, int num_pages
, int page_align
)
2746 struct ceph_osd_request
*req
;
2749 dout("readpages on ino %llx.%llx on %llu~%llu\n", vino
.ino
,
2750 vino
.snap
, off
, *plen
);
2751 req
= ceph_osdc_new_request(osdc
, layout
, vino
, off
, plen
, 0, 1,
2752 CEPH_OSD_OP_READ
, CEPH_OSD_FLAG_READ
,
2753 NULL
, truncate_seq
, truncate_size
,
2756 return PTR_ERR(req
);
2758 /* it may be a short read due to an object boundary */
2760 osd_req_op_extent_osd_data_pages(req
, 0,
2761 pages
, *plen
, page_align
, false, false);
2763 dout("readpages final extent is %llu~%llu (%llu bytes align %d)\n",
2764 off
, *plen
, *plen
, page_align
);
2766 ceph_osdc_build_request(req
, off
, NULL
, vino
.snap
, NULL
);
2768 rc
= ceph_osdc_start_request(osdc
, req
, false);
2770 rc
= ceph_osdc_wait_request(osdc
, req
);
2772 ceph_osdc_put_request(req
);
2773 dout("readpages result %d\n", rc
);
2776 EXPORT_SYMBOL(ceph_osdc_readpages
);
2779 * do a synchronous write on N pages
2781 int ceph_osdc_writepages(struct ceph_osd_client
*osdc
, struct ceph_vino vino
,
2782 struct ceph_file_layout
*layout
,
2783 struct ceph_snap_context
*snapc
,
2785 u32 truncate_seq
, u64 truncate_size
,
2786 struct timespec
*mtime
,
2787 struct page
**pages
, int num_pages
)
2789 struct ceph_osd_request
*req
;
2791 int page_align
= off
& ~PAGE_MASK
;
2793 BUG_ON(vino
.snap
!= CEPH_NOSNAP
); /* snapshots aren't writeable */
2794 req
= ceph_osdc_new_request(osdc
, layout
, vino
, off
, &len
, 0, 1,
2796 CEPH_OSD_FLAG_ONDISK
| CEPH_OSD_FLAG_WRITE
,
2797 snapc
, truncate_seq
, truncate_size
,
2800 return PTR_ERR(req
);
2802 /* it may be a short write due to an object boundary */
2803 osd_req_op_extent_osd_data_pages(req
, 0, pages
, len
, page_align
,
2805 dout("writepages %llu~%llu (%llu bytes)\n", off
, len
, len
);
2807 ceph_osdc_build_request(req
, off
, snapc
, CEPH_NOSNAP
, mtime
);
2809 rc
= ceph_osdc_start_request(osdc
, req
, true);
2811 rc
= ceph_osdc_wait_request(osdc
, req
);
2813 ceph_osdc_put_request(req
);
2816 dout("writepages result %d\n", rc
);
2819 EXPORT_SYMBOL(ceph_osdc_writepages
);
2821 int ceph_osdc_setup(void)
2823 size_t size
= sizeof(struct ceph_osd_request
) +
2824 CEPH_OSD_SLAB_OPS
* sizeof(struct ceph_osd_req_op
);
2826 BUG_ON(ceph_osd_request_cache
);
2827 ceph_osd_request_cache
= kmem_cache_create("ceph_osd_request", size
,
2830 return ceph_osd_request_cache
? 0 : -ENOMEM
;
2832 EXPORT_SYMBOL(ceph_osdc_setup
);
2834 void ceph_osdc_cleanup(void)
2836 BUG_ON(!ceph_osd_request_cache
);
2837 kmem_cache_destroy(ceph_osd_request_cache
);
2838 ceph_osd_request_cache
= NULL
;
2840 EXPORT_SYMBOL(ceph_osdc_cleanup
);
2843 * handle incoming message
2845 static void dispatch(struct ceph_connection
*con
, struct ceph_msg
*msg
)
2847 struct ceph_osd
*osd
= con
->private;
2848 struct ceph_osd_client
*osdc
;
2849 int type
= le16_to_cpu(msg
->hdr
.type
);
2856 case CEPH_MSG_OSD_MAP
:
2857 ceph_osdc_handle_map(osdc
, msg
);
2859 case CEPH_MSG_OSD_OPREPLY
:
2860 handle_reply(osdc
, msg
);
2862 case CEPH_MSG_WATCH_NOTIFY
:
2863 handle_watch_notify(osdc
, msg
);
2867 pr_err("received unknown message type %d %s\n", type
,
2868 ceph_msg_type_name(type
));
2875 * Lookup and return message for incoming reply. Don't try to do
2876 * anything about a larger than preallocated data portion of the
2877 * message at the moment - for now, just skip the message.
2879 static struct ceph_msg
*get_reply(struct ceph_connection
*con
,
2880 struct ceph_msg_header
*hdr
,
2883 struct ceph_osd
*osd
= con
->private;
2884 struct ceph_osd_client
*osdc
= osd
->o_osdc
;
2886 struct ceph_osd_request
*req
;
2887 int front_len
= le32_to_cpu(hdr
->front_len
);
2888 int data_len
= le32_to_cpu(hdr
->data_len
);
2891 tid
= le64_to_cpu(hdr
->tid
);
2892 mutex_lock(&osdc
->request_mutex
);
2893 req
= __lookup_request(osdc
, tid
);
2895 dout("%s osd%d tid %llu unknown, skipping\n", __func__
,
2902 ceph_msg_revoke_incoming(req
->r_reply
);
2904 if (front_len
> req
->r_reply
->front_alloc_len
) {
2905 pr_warn("%s osd%d tid %llu front %d > preallocated %d\n",
2906 __func__
, osd
->o_osd
, req
->r_tid
, front_len
,
2907 req
->r_reply
->front_alloc_len
);
2908 m
= ceph_msg_new(CEPH_MSG_OSD_OPREPLY
, front_len
, GFP_NOFS
,
2912 ceph_msg_put(req
->r_reply
);
2916 if (data_len
> req
->r_reply
->data_length
) {
2917 pr_warn("%s osd%d tid %llu data %d > preallocated %zu, skipping\n",
2918 __func__
, osd
->o_osd
, req
->r_tid
, data_len
,
2919 req
->r_reply
->data_length
);
2925 m
= ceph_msg_get(req
->r_reply
);
2926 dout("get_reply tid %lld %p\n", tid
, m
);
2929 mutex_unlock(&osdc
->request_mutex
);
2933 static struct ceph_msg
*alloc_msg(struct ceph_connection
*con
,
2934 struct ceph_msg_header
*hdr
,
2937 struct ceph_osd
*osd
= con
->private;
2938 int type
= le16_to_cpu(hdr
->type
);
2939 int front
= le32_to_cpu(hdr
->front_len
);
2943 case CEPH_MSG_OSD_MAP
:
2944 case CEPH_MSG_WATCH_NOTIFY
:
2945 return ceph_msg_new(type
, front
, GFP_NOFS
, false);
2946 case CEPH_MSG_OSD_OPREPLY
:
2947 return get_reply(con
, hdr
, skip
);
2949 pr_info("alloc_msg unexpected msg type %d from osd%d\n", type
,
2957 * Wrappers to refcount containing ceph_osd struct
2959 static struct ceph_connection
*get_osd_con(struct ceph_connection
*con
)
2961 struct ceph_osd
*osd
= con
->private;
2967 static void put_osd_con(struct ceph_connection
*con
)
2969 struct ceph_osd
*osd
= con
->private;
2977 * Note: returned pointer is the address of a structure that's
2978 * managed separately. Caller must *not* attempt to free it.
2980 static struct ceph_auth_handshake
*get_authorizer(struct ceph_connection
*con
,
2981 int *proto
, int force_new
)
2983 struct ceph_osd
*o
= con
->private;
2984 struct ceph_osd_client
*osdc
= o
->o_osdc
;
2985 struct ceph_auth_client
*ac
= osdc
->client
->monc
.auth
;
2986 struct ceph_auth_handshake
*auth
= &o
->o_auth
;
2988 if (force_new
&& auth
->authorizer
) {
2989 ceph_auth_destroy_authorizer(auth
->authorizer
);
2990 auth
->authorizer
= NULL
;
2992 if (!auth
->authorizer
) {
2993 int ret
= ceph_auth_create_authorizer(ac
, CEPH_ENTITY_TYPE_OSD
,
2996 return ERR_PTR(ret
);
2998 int ret
= ceph_auth_update_authorizer(ac
, CEPH_ENTITY_TYPE_OSD
,
3001 return ERR_PTR(ret
);
3003 *proto
= ac
->protocol
;
3009 static int verify_authorizer_reply(struct ceph_connection
*con
, int len
)
3011 struct ceph_osd
*o
= con
->private;
3012 struct ceph_osd_client
*osdc
= o
->o_osdc
;
3013 struct ceph_auth_client
*ac
= osdc
->client
->monc
.auth
;
3015 return ceph_auth_verify_authorizer_reply(ac
, o
->o_auth
.authorizer
, len
);
3018 static int invalidate_authorizer(struct ceph_connection
*con
)
3020 struct ceph_osd
*o
= con
->private;
3021 struct ceph_osd_client
*osdc
= o
->o_osdc
;
3022 struct ceph_auth_client
*ac
= osdc
->client
->monc
.auth
;
3024 ceph_auth_invalidate_authorizer(ac
, CEPH_ENTITY_TYPE_OSD
);
3025 return ceph_monc_validate_auth(&osdc
->client
->monc
);
3028 static int osd_sign_message(struct ceph_msg
*msg
)
3030 struct ceph_osd
*o
= msg
->con
->private;
3031 struct ceph_auth_handshake
*auth
= &o
->o_auth
;
3033 return ceph_auth_sign_message(auth
, msg
);
3036 static int osd_check_message_signature(struct ceph_msg
*msg
)
3038 struct ceph_osd
*o
= msg
->con
->private;
3039 struct ceph_auth_handshake
*auth
= &o
->o_auth
;
3041 return ceph_auth_check_message_signature(auth
, msg
);
3044 static const struct ceph_connection_operations osd_con_ops
= {
3047 .dispatch
= dispatch
,
3048 .get_authorizer
= get_authorizer
,
3049 .verify_authorizer_reply
= verify_authorizer_reply
,
3050 .invalidate_authorizer
= invalidate_authorizer
,
3051 .alloc_msg
= alloc_msg
,
3052 .sign_message
= osd_sign_message
,
3053 .check_message_signature
= osd_check_message_signature
,