forked from scality/RestBlockDriver
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsrb_driver.c
1712 lines (1446 loc) · 41 KB
/
srb_driver.c
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/*
* Copyright (C) 2014 SCALITY SA - http://www.scality.com
*
* This file is part of ScalityRestBlock.
*
* ScalityRestBlock is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* ScalityRestBlock is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with ScalityRestBlock. If not, see <http://www.gnu.org/licenses/>.
*
*/
// Defining __KERNEL__ and MODULE allows us to access kernel-level
// code not usually available to userspace programs.
#undef __KERNEL__
#define __KERNEL__
#undef MODULE
#define MODULE
// Linux Kernel/LKM headers: module.h is needed by all modules and
// kernel.h is needed for KERN_INFO.
#include <linux/module.h> // included for all kernel modules
#include <linux/kernel.h> // included for KERN_INFO
#include <linux/moduleparam.h> // included for LKM parameters
#include <linux/init.h> // included for __init and __exit macros
#include <linux/device.h>
#include <linux/spinlock.h>
#include <linux/wait.h>
#include <linux/mutex.h>
#include <linux/fs.h>
#include <linux/blkdev.h>
#include <linux/slab.h>
#include <linux/kthread.h>
#include <linux/blkdev.h>
#include "srb.h"
/* TODO: Testing REQ_FLUSH with old version (Issue #21)
*/
//#define SRB_BIO_ENABLED 1
// LKM information
MODULE_AUTHOR("Laurent Meyer <[email protected]>");
MODULE_AUTHOR("David Pineau <[email protected]>");
MODULE_DESCRIPTION("Block Device Driver for REST-based storage");
MODULE_LICENSE("GPL");
MODULE_VERSION(DEV_REL_VERSION);
static srb_device_t devtab[DEV_MAX];
static srb_server_t *servers = NULL;
static DEFINE_SPINLOCK(devtab_lock);
/* Module parameters (LKM parameters)
*/
unsigned short srb_log = SRB_LOG_LEVEL_DFLT;
unsigned short req_timeout = SRB_REQ_TIMEOUT_DFLT;
unsigned short nb_req_retries = SRB_NB_REQ_RETRIES_DFLT;
unsigned short server_conn_timeout = SRB_CONN_TIMEOUT_DFLT;
unsigned int thread_pool_size = SRB_THREAD_POOL_SIZE_DFLT;
MODULE_PARM_DESC(debug, "Global log level for ScalityRestBlock LKM");
module_param_named(debug, srb_log, ushort, 0644);
MODULE_PARM_DESC(req_timeout, "Global timeout for request");
module_param(req_timeout, ushort, 0644);
MODULE_PARM_DESC(nb_req_retries, "Global number of retries for request");
module_param(nb_req_retries, ushort, 0644);
MODULE_PARM_DESC(server_conn_timeout, "Global timeout for connection to server(s)");
module_param(server_conn_timeout, ushort, 0644);
MODULE_PARM_DESC(thread_pool_size, "Size of the thread pool");
module_param(thread_pool_size, uint, 0444);
/* XXX: Request mapping
*/
char *req_code_to_str(int code)
{
switch (code) {
case READ: return "READ"; break;
case WRITE: return "WRITE"; break;
case WRITE_FLUSH: return "WRITE_FLUSH"; break;
case WRITE_FUA: return "WRITE_FUA"; break;
case WRITE_FLUSH_FUA: return "WRITE_FLUSH_FUA"; break;
default: return "UNKNOWN";
}
}
int req_flags_to_str(int flags, char *buff)
{
//char buff[128];
int size = 0;
buff[0] = '\0';
// detect common flags
if (flags == REQ_COMMON_MASK) {
strncpy(buff, "REQ_COMMON_MASK", 15);
buff[16] = '\0';
return 16;
}
if (flags == REQ_FAILFAST_MASK) {
strncpy(buff, "REQ_FAILFAST_MASK", 17);
buff[18] = '\0';
return 18;
}
if (flags == REQ_NOMERGE_FLAGS) {
strncpy(buff, "REQ_NOMERGE_FLAGS", 17);
buff[18] = '\0';
return 18;
}
if (flags & REQ_WRITE) {
strncpy(&buff[size], "REQ_WRITE|", 10);
size += 10;
}
if (flags & REQ_FAILFAST_DEV) {
strncpy(&buff[size], "REQ_FAILFAST_DEV|", 17);
size += 17;
}
if (flags & REQ_FAILFAST_TRANSPORT) {
strncpy(&buff[size], "REQ_FAILFAST_TRANSPORT|", 23);
size += 23;
}
if (flags & REQ_FAILFAST_DRIVER) {
strncpy(&buff[size], "REQ_FAILFAST_DRIVER|", 20);
size += 20;
}
if (flags & REQ_SYNC) {
strncpy(&buff[size], "REQ_SYNC|", 9);
size += 9;
}
if (flags & REQ_META) {
strncpy(&buff[size], "REQ_META|", 9);
size += 9;
}
if (flags & REQ_PRIO) {
strncpy(&buff[size], "REQ_PRIO|", 9);
size += 9;
}
if (flags & REQ_DISCARD) {
strncpy(&buff[size], "REQ_DISCARD|", 13);
size += 13;
}
if (flags & REQ_WRITE_SAME) {
strncpy(&buff[size], "REQ_WRITE_SAME|", 16);
size += 16;
}
if (flags & REQ_NOIDLE) {
strncpy(&buff[size], "REQ_NOIDLE|", 11);
size += 11;
}
if (flags & REQ_RAHEAD) {
strncpy(&buff[size], "REQ_RAHEAD|", 11);
size += 11;
}
if (flags & REQ_THROTTLED) {
strncpy(&buff[size], "REQ_THROTTLED|", 14);
size += 14;
}
if (flags & REQ_SORTED) {
strncpy(&buff[size], "REQ_SORTED|", 11);
size += 11;
}
if (flags & REQ_SOFTBARRIER) {
strncpy(&buff[size], "REQ_SOFTBARRIER|", 16);
size += 16;
}
if (flags & REQ_FUA) {
strncpy(&buff[size], "REQ_FUA|", 8);
size += 8;
}
if (flags & REQ_NOMERGE) {
strncpy(&buff[size], "REQ_NOMERGE|", 12);
size += 12;
}
if (flags & REQ_STARTED) {
strncpy(&buff[size], "REQ_STARTED|", 12);
size += 12;
}
if (flags & REQ_DONTPREP) {
strncpy(&buff[size], "REQ_DONTPREP|", 13);
size += 13;
}
if (flags & REQ_QUEUED) {
strncpy(&buff[size], "REQ_QUEUED|", 11);
size += 11;
}
if (flags & REQ_ELVPRIV) {
strncpy(&buff[size], "REQ_ELVPRIV|", 12);
size += 12;
}
if (flags & REQ_FAILED) {
strncpy(&buff[size], "REQ_FAILED|", 11);
size += 11;
}
if (flags & REQ_QUIET) {
strncpy(&buff[size], "REQ_QUIET|", 10);
size += 10;
}
if (flags & REQ_PREEMPT) {
strncpy(&buff[size], "REQ_PREEMPT|", 12);
size += 12;
}
if (flags & REQ_ALLOCED) {
strncpy(&buff[size], "REQ_ALLOCED|", 12);
size += 12;
}
if (flags & REQ_COPY_USER) {
strncpy(&buff[size], "REQ_COPY_USER|", 14);
size += 14;
}
if (flags & REQ_FLUSH) {
strncpy(&buff[size], "REQ_FLUSH|", 10);
size += 10;
}
if (flags & REQ_FLUSH_SEQ) {
strncpy(&buff[size], "REQ_FLUSH_SEQ|", 14);
size += 14;
}
if (flags & REQ_IO_STAT) {
strncpy(&buff[size], "REQ_IO_STAT|", 12);
size += 12;
}
if (flags & REQ_MIXED_MERGE) {
strncpy(&buff[size], "REQ_MIXED_MERGE|", 16);
size += 16;
}
if (flags & REQ_SECURE) {
strncpy(&buff[size], "REQ_SECURE|", 11);
size += 11;
}
if (flags & REQ_KERNEL) {
strncpy(&buff[size], "REQ_KERNEL|", 11);
size += 11;
}
if (flags & REQ_PM) {
strncpy(&buff[size], "REQ_PM|", 7);
size += 7;
}
#ifdef REQ_END /* appears in 3.6.10 in linux/blk_types.h */
if (flags & REQ_END) {
strncpy(&buff[size], "REQ_END|", 8);
size += 8;
}
#endif
if (size != 0)
buff[size-1] = '\0';
return size;
}
/*
* Handle an I/O request.
*/
#ifdef SRB_BIO_ENABLED
static void srb_xmit_range(struct srb_device_s *dev,
struct srb_cdmi_desc_s *desc,
char *buf,
unsigned long range_start, unsigned long size, int write)
{
if ((range_start + size) > dev->disk_size) {
SRB_INFO("Beyond-end write (%lu %lu)", range_start, size);
return;
}
if (size != 4096)
SRB_DEV_DEBUG("Wrote size of %lu", size);
if (write) {
srb_cdmi_putrange(&dev->debug,
desc,
range_start,
size);
}
else {
srb_cdmi_getrange(&dev->debug,
desc,
//buf,
range_start,
size);
}
}
static int srb_xfer_bio(struct srb_device_s *dev,
struct srb_cdmi_desc_s *desc,
struct bio *bio)
{
int i;
struct bio_vec *bvec;
sector_t sector = bio->bi_sector;
/*
** Takes care of REQ_FLUSH flag, if set we have to flush
** before sending actual bio to device
*/
if (bio->bi_rw & REQ_FLUSH) {
SRB_DEV_DEBUG("[Flush(REG_FLUSH)]");
srb_cdmi_flush(&dev->debug, desc, dev->disk_size);
}
bio_for_each_segment(bvec, bio, i) {
char *buffer = kmap(bvec->bv_page);
unsigned int nbsect = bvec->bv_len / 512UL;
SRB_DEV_DEBUG("[Transfering sect=%lu, nb=%d w=%d]",
(unsigned long) sector, nbsect,
bio_data_dir(bio) == WRITE);
srb_xmit_range(dev, desc, buffer + bvec->bv_offset,
sector * 512UL, nbsect * 512UL,
bio_data_dir(bio) == WRITE);
sector += nbsect;
kunmap(bvec->bv_page);
}
/*
** Takes care care of REQ_FUA flag, need to flush here
** after the bio have been sent
*/
if (bio->bi_rw & REQ_FUA) {
SRB_DEV_DEBUG("[Flush(REG_FUA)]");
srb_cdmi_flush(&dev->debug, desc, dev->disk_size);
}
return 0;
}
#endif /* _SRB_BIO_ENABLED_ */
/* TODO: Handle CDMI request timeout and retry (Issue #22)
* For error handling use a returned code
*/
//void srb_xfer_scl(struct srb_device_s *dev,
int srb_xfer_scl(struct srb_device_s *dev,
struct srb_cdmi_desc_s *desc,
struct request *req)
{
int ret;
int i;
struct timeval tv_start;
struct timeval tv_end;
SRB_LOG_DEBUG(dev->debug.level, "srb_xfer_scl: CDMI request %p (%s) for device %p with cdmi %p",
req, req_code_to_str(rq_data_dir(req)), dev, desc);
ret = 0;
/* TODO: Handle CDMI request retry (Issue #22)
*/
if (SRB_DEBUG <= dev->debug.level)
do_gettimeofday(&tv_start);
for (i = 0; i < nb_req_retries; i++) {
if (rq_data_dir(req) == WRITE) {
ret = srb_cdmi_putrange(&dev->debug,
desc,
blk_rq_pos(req) * 512ULL,
blk_rq_sectors(req) * 512ULL);
}
else {
ret = srb_cdmi_getrange(&dev->debug,
desc,
blk_rq_pos(req) * 512ULL,
blk_rq_sectors(req) * 512ULL);
}
if (0 == ret)
break;
else if (i < nb_req_retries - 1)
SRB_LOG_NOTICE(dev->debug.level, "Retrying CDMI request... %d", (i + 1));
}
if (SRB_DEBUG <= dev->debug.level) {
do_gettimeofday(&tv_end);
SRB_LOG_DEBUG(dev->debug.level, "cdmi request time: %ldms",
(tv_end.tv_sec - tv_start.tv_sec)*1000 + (tv_end.tv_usec - tv_start.tv_usec)/1000);
}
if (ret) {
SRB_LOG_ERR(dev->debug.level, "CDMI Request using scatterlist failed with IO error: %d", ret);
return -EIO;
}
return ret;
}
/*
* Free internal disk
*/
static int srb_free_disk(struct srb_device_s *dev)
{
struct gendisk *disk = NULL;
disk = dev->disk;
if (!disk) {
SRB_LOG_ERR(dev->debug.level, "%s: Disk is no more available", dev->name);
return -EINVAL;
}
dev->disk = NULL;
/* free disk */
if (disk->flags & GENHD_FL_UP) {
del_gendisk(disk);
if (disk->queue)
blk_cleanup_queue(disk->queue);
}
put_disk(disk);
return 0;
}
/*
* Thread for srb
*/
static int srb_thread(void *data)
{
struct srb_device_s *dev;
struct request *req;
unsigned long flags;
int th_id;
int th_ret = 0;
char buff[256];
#ifdef SRB_BIO_ENABLED
struct bio *bio;
#endif
struct req_iterator iter;
struct bio_vec *bvec;
struct srb_cdmi_desc_s *cdmi_desc;
SRB_LOG_DEBUG(((struct srb_device_s *)data)->debug.level, "srb_thread: thread function with data %p", data);
dev = data;
/* Init thread specific values */
spin_lock(&devtab_lock);
th_id = dev->nb_threads;
dev->nb_threads++;
spin_unlock(&devtab_lock);
set_user_nice(current, -20);
while (!kthread_should_stop() || !list_empty(&dev->waiting_queue)) {
/* wait for something to do */
wait_event_interruptible(dev->waiting_wq,
kthread_should_stop() ||
!list_empty(&dev->waiting_queue));
/* TODO: improve kthread termination, otherwise calling we can not
terminate a kthread calling kthread_stop() */
/* if (kthread_should_stop()) {
printk(KERN_INFO "srb_thread: immediate kthread exit\n");
do_exit(0);
} */
spin_lock_irqsave(&dev->waiting_lock, flags);
/* extract request */
if (list_empty(&dev->waiting_queue)) {
spin_unlock_irqrestore(&dev->waiting_lock, flags);
continue;
}
req = list_entry(dev->waiting_queue.next, struct request,
queuelist);
list_del_init(&req->queuelist);
spin_unlock_irqrestore(&dev->waiting_lock, flags);
if (blk_rq_sectors(req) == 0)
continue;
req_flags_to_str(req->cmd_flags, buff);
SRB_LOG_DEBUG(dev->debug.level, "srb_thread: thread %d: New REQ of type %s (%d) flags: %s (%llu)",
th_id, req_code_to_str(rq_data_dir(req)), rq_data_dir(req), buff, req->cmd_flags);
if (req->cmd_flags & REQ_FLUSH) {
SRB_LOG_DEBUG(dev->debug.level, "DEBUG CMD REQ_FLUSH\n");
}
/* XXX: Use iterator instead of internal function (cf linux/blkdev.h)
* __rq_for_each_bio(bio, req) {
*/
rq_for_each_segment(bvec, req, iter) {
if (iter.bio->bi_rw & REQ_FLUSH) {
SRB_LOG_DEBUG(dev->debug.level, "DEBUG VR BIO REQ_FLUSH\n");
}
}
#ifdef SRB_BIO_ENABLED
__rq_for_each_bio(bio, req) {
//srb_xfer_bio(dev, &dev->thread_cdmi_desc[th_id], bio);
srb_xfer_bio(dev, dev->thread_cdmi_desc[th_id], bio);
}
#else
/* Create scatterlist */
/* sg_init_table(dev->thread_cdmi_desc[th_id].sgl, DEV_NB_PHYS_SEGS);
dev->thread_cdmi_desc[th_id].sgl_size = blk_rq_map_sg(dev->q, req, dev->thread_cdmi_desc[th_id].sgl); */
cdmi_desc = dev->thread_cdmi_desc[th_id];
sg_init_table(dev->thread_cdmi_desc[th_id]->sgl, DEV_NB_PHYS_SEGS);
dev->thread_cdmi_desc[th_id]->sgl_size = blk_rq_map_sg(dev->q, req, dev->thread_cdmi_desc[th_id]->sgl);
SRB_LOG_DEBUG(dev->debug.level, "srb_thread: scatter_list size %d [nb_seg = %d, sector = %lu, nr_sectors=%u w=%d]",
DEV_NB_PHYS_SEGS, dev->thread_cdmi_desc[th_id]->sgl_size, blk_rq_pos(req), blk_rq_sectors(req), rq_data_dir(req) == WRITE);
/* Call scatter function */
//th_ret = srb_xfer_scl(dev, &dev->thread_cdmi_desc[th_id], req);
th_ret = srb_xfer_scl(dev, dev->thread_cdmi_desc[th_id], req);
#endif /* _SRB_BIO_ENABLED_ */
//SRB_DEV_DEBUG("END REQUEST [tid:%d]", th_id);
SRB_LOG_DEBUG(dev->debug.level, "srb_thread: thread %d: REQ done with returned code %d",
th_id, th_ret);;
/* No IO error testing for the moment */
blk_end_request_all(req, 0);
}
return 0;
}
static void srb_rq_fn(struct request_queue *q)
{
struct srb_device_s *dev = q->queuedata;
struct request *req;
unsigned long flags;
while ((req = blk_fetch_request(q)) != NULL) {
if (req->cmd_type != REQ_TYPE_FS) {
SRB_LOG_DEBUG(dev->debug.level, "srb_rq_fn: Skip non-CMD request");
__blk_end_request_all(req, -EIO);
continue;
}
spin_lock_irqsave(&dev->waiting_lock, flags);
list_add_tail(&req->queuelist, &dev->waiting_queue);
spin_unlock_irqrestore(&dev->waiting_lock, flags);
wake_up_nr(&dev->waiting_wq, 1);
}
}
static int srb_open(struct block_device *bdev, fmode_t mode)
{
//srb_device_t *dev = bdev->bd_disk->private_data;
srb_device_t *dev;
SRB_LOG_INFO(srb_log, "srb_open: opening block device %s", bdev->bd_disk->disk_name);
dev = bdev->bd_disk->private_data;
spin_lock(&devtab_lock);
dev->users++;
spin_unlock(&devtab_lock);
return 0;
}
/*
* After linux kernel v3.10, this function stops returning anything
* (becomes void). For simplicity, we currently don't support earlier kernels.
*/
static void srb_release(struct gendisk *disk, fmode_t mode)
{
srb_device_t *dev;
SRB_LOG_INFO(srb_log, "srb_release: releasing disk %s", disk->disk_name);
dev = disk->private_data;
spin_lock(&devtab_lock);
dev->users--;
spin_unlock(&devtab_lock);
}
static const struct block_device_operations srb_fops =
{
.owner = THIS_MODULE,
.open = srb_open,
.release = srb_release,
};
static int srb_init_disk(struct srb_device_s *dev)
{
struct gendisk *disk = NULL;
struct request_queue *q;
int i;
int ret = 0;
SRB_LOG_INFO(srb_log, "srb_init_disk: initializing disk for device: %s", dev->name);
/* create gendisk info */
disk = alloc_disk(DEV_MINORS);
if (!disk) {
SRB_LOG_WARN(srb_log, "srb_init_disk: unable to allocate memory for disk for device: %s",
dev->name);
return -ENOMEM;
}
SRB_LOG_DEBUG(srb_log, "Creating new disk: %p", disk);
strcpy(disk->disk_name, dev->name);
disk->major = dev->major;
disk->first_minor = 0;
disk->fops = &srb_fops;
disk->private_data = dev;
/* init rq */
q = blk_init_queue(srb_rq_fn, &dev->rq_lock);
if (!q) {
SRB_LOG_WARN(srb_log, "srb_init_disk: unable to init block queue for device: %p, disk: %p",
dev, disk);
//put_disk(disk);
srb_free_disk(dev);
return -ENOMEM;
}
blk_queue_max_hw_sectors(q, DEV_NB_PHYS_SEGS);
q->queuedata = dev;
dev->disk = disk;
dev->q = disk->queue = q;
dev->nb_threads = 0;
//blk_queue_flush(q, REQ_FLUSH | REQ_FUA);
//blk_queue_max_phys_segments(q, DEV_NB_PHYS_SEGS);
//TODO: Enable flush and bio (Issue #21)
//blk_queue_flush(q, REQ_FLUSH);
//TODO: Make thread pool variable (Issue #33)
//for (i = 0; i < SRB_THREAD_POOL_SIZE; i++) {
for (i = 0; i < thread_pool_size; i++) {
//if ((ret = srb_cdmi_connect(&dev->debug, &dev->thread_cdmi_desc[i]))) {
if ((ret = srb_cdmi_connect(&dev->debug, dev->thread_cdmi_desc[i]))) {
SRB_LOG_ERR(srb_log, "Unable to connect to CDMI endpoint: %d",
ret);
//put_disk(disk);
srb_free_disk(dev);
return -EIO;
}
}
/* Caution: be sure to call this before spawning threads */
//ret = srb_cdmi_getsize(&dev->debug, &dev->thread_cdmi_desc[0],
ret = srb_cdmi_getsize(&dev->debug, dev->thread_cdmi_desc[0], &dev->disk_size);
if (ret != 0) {
SRB_LOG_ERR(srb_log, "Could not retrieve volume size.");
//put_disk(disk);
srb_free_disk(dev);
return ret;
}
set_capacity(disk, dev->disk_size / 512ULL);
//TODO: Make thread pool variable (Issue #33)
//for (i = 0; i < SRB_THREAD_POOL_SIZE; i++) {
for (i = 0; i < thread_pool_size; i++) {
dev->thread[i] = kthread_create(srb_thread, dev, "%s",
dev->disk->disk_name);
if (IS_ERR(dev->thread[i])) {
SRB_LOG_ERR(srb_log, "Unable to create worker thread (id %d)", i);
//put_disk(disk);
dev->thread[i] = NULL;
srb_free_disk(dev);
goto err_kthread;
}
wake_up_process(dev->thread[i]);
}
add_disk(disk);
SRB_LOG_INFO(srb_log, "%s: Added of size 0x%llx",
disk->disk_name, (unsigned long long)dev->disk_size);
return 0;
err_kthread:
for (i = 0; i < thread_pool_size; i++) {
if (dev->thread[i] != NULL)
kthread_stop(dev->thread[i]);
}
return -EIO;
}
#define device_free_slot(X) ((X)->name[0] == 0)
/* This function gets the next free slot in device tab (devtab)
** and sets its name and id.
** Note : all the remaining fields states are undefived, it is
** the caller responsability to set them.
*/
static int srb_device_new(const char *devname, srb_device_t *dev)
{
int ret = -EINVAL;
int i;
SRB_LOG_INFO(srb_log, "srb_device_new: creating new device %s"
" with %d threads", devname, thread_pool_size);
if (NULL == dev) {
ret = -EINVAL;
goto out;
}
if (NULL == devname || strlen(devname) >= DISK_NAME_LEN) {
SRB_LOG_ERR(srb_log, "srb_device_new: "
"Invalid (or too long) device name '%s'",
devname == NULL ? "" : devname);
ret = -EINVAL;
goto out;
}
/* Lock table to protect against concurrent devices
* creation
*/
dev->debug.name = &dev->name[0];
dev->debug.level = srb_log;
dev->users = 0;
strncpy(dev->name, devname, strlen(devname));
/* XXX: dynamic allocation of thread pool and cdmi connection pool
* NB: The memory allocation for the thread is an array of pointer
* whereas the allocation for the cdmi connection pool is an array
* of cdmi connection structure
*/
dev->thread_cdmi_desc = kmalloc(sizeof(struct srb_cdmi_desc_s *) * thread_pool_size, GFP_KERNEL);
if (dev->thread_cdmi_desc == NULL) {
SRB_LOG_CRIT(srb_log, "srb_device_new: Unable to allocate memory for CDMI struct pointer");
ret = -ENOMEM;
goto err_mem;
}
for (i = 0; i < thread_pool_size; i++) {
dev->thread_cdmi_desc[i] = kmalloc(sizeof(struct srb_cdmi_desc_s), GFP_KERNEL);
if (dev->thread_cdmi_desc[i] == NULL) {
SRB_LOG_CRIT(srb_log, "srb_device_new: Unable to allocate memory for CDMI struct, step %d", i);
ret = -ENOMEM;
goto err_mem;
}
}
dev->thread = kmalloc(sizeof(struct task_struct *) * thread_pool_size, GFP_KERNEL);
if (dev->thread == NULL) {
SRB_LOG_CRIT(srb_log, "srb_device_new: Unable to allocate memory for kernel thread struct");
ret = -ENOMEM;
goto err_mem;
}
return 0;
err_mem:
if (NULL != dev && NULL != dev->thread_cdmi_desc) {
for (i = 0; i < thread_pool_size; i++) {
if (dev->thread_cdmi_desc[i])
kfree(dev->thread_cdmi_desc[i]);
}
kfree(dev->thread_cdmi_desc);
}
out:
return ret;
}
/* This helper marks the given device slot as empty
** CAUTION: the devab lock must be held
*/
static void __srb_device_free(srb_device_t *dev)
{
SRB_LOG_INFO(srb_log, "__srb_device_free: freeing device: %s", dev->name);
memset(dev->name, 0, DISK_NAME_LEN);
dev->major = 0;
dev->id = -1;
}
static void srb_device_free(srb_device_t *dev)
{
int i = 0;
SRB_LOG_INFO(srb_log, "srb_device_free: freeing device: %s", dev->name);
__srb_device_free(dev);
/* TODO: free kernel memory for CDMI struct (Issue #33)
*/
if (dev->thread_cdmi_desc) {
for (i = 0; i < thread_pool_size; i++) {
if (dev->thread_cdmi_desc[i])
kfree(dev->thread_cdmi_desc[i]);
}
kfree(dev->thread_cdmi_desc);
}
if (dev->thread)
kfree(dev->thread);
}
static int _srb_reconstruct_url(char *url, char *name,
const char *baseurl, const char *basepath,
const char *filename)
{
int urllen = 0;
int namelen = 0;
int seplen = 0;
SRB_LOG_DEBUG(srb_log, "_srb_reconstruct_url: construction of URL with url: %s, name: %s, baseurl: %s, basepath: %p, filename: %s",
url, name, baseurl, basepath, filename);
urllen = strlen(baseurl);
if (baseurl[urllen - 1] != '/')
seplen = 1;
if (seplen)
{
urllen = snprintf(url, SRB_URL_SIZE, "%s/%s", baseurl, filename);
namelen = snprintf(name, SRB_URL_SIZE, "%s/%s", basepath, filename);
}
else
{
urllen = snprintf(url, SRB_URL_SIZE, "%s%s", baseurl, filename);
namelen = snprintf(name, SRB_URL_SIZE, "%s%s", basepath, filename);
}
if (urllen >= SRB_URL_SIZE || namelen >= SRB_URL_SIZE)
return -EINVAL;
return 0;
}
static int __srb_device_detach(srb_device_t *dev)
{
int i;
int ret = 0;
SRB_LOG_DEBUG(srb_log, "__srb_device_detach: detaching device %s (%p)", dev->name, dev);
if (!dev) {
SRB_LOG_WARN(srb_log, "__srb_device_detach: empty device");
return -EINVAL;
}
if (dev->users > 0) {
SRB_LOG_ERR(srb_log, "%s: Unable to remove, device still opened (#users: %d)", dev->name, dev->users);
return -EBUSY;
}
if (!dev->disk) {
SRB_LOG_ERR(srb_log, "%s: Disk is no more available", dev->name);
return -EINVAL;
}
SRB_LOG_INFO(srb_log, "%s: Removing disk", dev->disk->disk_name);
/* TODO: Make thread pool variable (Issue #33)
*/
for (i = 0; i < thread_pool_size; i++) {
if (dev->thread[i])
kthread_stop(dev->thread[i]);
}
/* free disk */
ret = srb_free_disk(dev);
if (0 != ret) {
SRB_LOG_WARN(srb_log, "%s: Failed to remove disk: %d", dev->name, ret);
}
SRB_LOG_INFO(srb_log, "%s: Removing disk for major %d", dev->name, dev->major);
/* Remove device */
unregister_blkdev(dev->major, DEV_NAME);
/* Mark slot as empty */
if (NULL != dev)
srb_device_free(dev);
return 0;
}
static int _srb_detach_devices(void)
{
int ret;
int i = 0;
int errcount = 0;
int dev_in_use[DEV_MAX];
SRB_LOG_INFO(srb_log, "_srb_detach_devices: detaching devices");
/* mark all device that are not used */
spin_lock(&devtab_lock);
for (i = 0; i < DEV_MAX; ++i) {
dev_in_use[i] = 0;
if (!device_free_slot(&devtab[i])) {
if (devtab[i].state == DEV_IN_USE) {
dev_in_use[i] = 1;
} else {
devtab[i].state = DEV_IN_USE;
dev_in_use[i] = 2;
}
}
}
spin_unlock(&devtab_lock);
/* detach all marked devices */
for (i = 0; i < DEV_MAX; ++i) {
if (2 == dev_in_use[i]) {
ret = __srb_device_detach(&devtab[i]);
if (ret != 0) {
SRB_LOG_ERR(srb_log, "Could not remove device %s for volume at unload %s",
devtab[i].name, devtab[i].thread_cdmi_desc ? devtab[i].thread_cdmi_desc[0]->filename : "NULL");
errcount++;
}
}
}
/* mark all device that are not used */
spin_lock(&devtab_lock);
for (i = 0; i < DEV_MAX; ++i) {
if (2 == dev_in_use[i]) {
devtab[i].state = DEV_UNUSED;
}
}
spin_unlock(&devtab_lock);
return errcount;
}
static void _srb_server_free(srb_server_t *server)
{
SRB_LOG_DEBUG(srb_log, "_srb_server_free: deleting server url %s (%p)", server->cdmi_desc.url, server);
if (server) {
server->next = NULL;
kfree(server);
server = NULL;
}
}
static int _srb_server_new(srb_debug_t *dbg, const char *url, srb_server_t **server)
{
srb_server_t *new = NULL;
int ret = 0;
SRB_LOG_DEBUG(dbg->level, "_srb_server_new: creating server with url: %s, servers: %p", url, *server);
new = kcalloc(1, sizeof(struct srb_server_s), GFP_KERNEL);
if (new == NULL) {
SRB_LOG_ERR(dbg->level, "Cannot allocate memory to add a new server.");
ret = -ENOMEM;
goto end;
}
ret = srb_cdmi_init(dbg, &new->cdmi_desc, url);
if (ret != 0) {
SRB_LOG_ERR(dbg->level, "Could not initialize server descriptor (parse URL).");
goto end;
}
if (server) {
new->next = NULL;
*server = new;
}
return 0;
end:
_srb_server_free(new);
return ret;
}
/*
* XXX NOTE XXX: #13 This function picks only one server that has enough free
* space in the URL buffer to append the filename.
*/
static int _srb_server_pick(const char *filename, struct srb_cdmi_desc_s *pick)
{
char url[SRB_URL_SIZE];
char name[SRB_URL_SIZE];
int ret;
int found = 0;
srb_server_t *server = NULL;
SRB_LOG_DEBUG(srb_log, "_srb_server_pick: picking server with filename: %s, with CDMI pick %p", filename, pick);
spin_lock(&devtab_lock);
server = servers;
while (server != NULL) {
SRB_LOG_INFO(srb_log, "Browsing server: %s", server->cdmi_desc.url);
ret = _srb_reconstruct_url(url, name,
server->cdmi_desc.url,
server->cdmi_desc.filename,
filename);
SRB_LOG_INFO(srb_log, "Dewb reconstruct url yielded %s, %i", url, ret);
if (ret == 0) {
memcpy(pick, &server->cdmi_desc, sizeof(struct srb_cdmi_desc_s));
strncpy(pick->url, url, SRB_URL_SIZE);
strncpy(pick->filename, name, SRB_URL_SIZE);
SRB_LOG_INFO(srb_log, "Copied into pick: url=%s, name=%s", pick->url, pick->filename);
found = 1;
break ;
}