summaryrefslogtreecommitdiff
path: root/mysys/raid.cc
blob: 29819a878c4bf5f71de6f45bfcf0972a9a04333d (plain)
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
/* Copyright (C) 2000 MySQL AB

   This program 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 2 of the License, or
   (at your option) any later version.

   This program 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 this program; if not, write to the Free Software
   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA */

/*
 
   RAID support for MySQL. Raid 0 (stiping) only implemented yet.
 
   Why RAID? Why it must be in MySQL?
 
   This is because then you can:
   1. Have bigger tables than your OS limit. In time of writing this
      we are hitting to 2GB limit under linux/ext2
   2. You can get more speed from IO bottleneck by putting
      Raid dirs on different physical disks.
   3. Getting more fault tolerance (not implemented yet)
 
   Why not to use RAID:
 
   1. You are losing some processor power to calculate things,
      do more syscalls and interrupts.
 
   Functionality is supplied by two classes: RaidFd and RaidName.
   RaidFd supports funtionality over file descriptors like
   open/create/write/seek/close. RaidName supports functionality
   like rename/delete where we have no relations to filedescriptors.
   RaidName can be prorably unchanged for different Raid levels. RaidFd
   have to be virtual I think ;).
   You can speed up some calls in MySQL code by skipping RAID code.
   For example LOAD DATA INFILE never needs to read RAID-ed files.
   This can be done adding proper "#undef my_read" or similar undef-s
   in your code. Check out the raid.h!
 
   Some explanation about _seek_vector[]
   This is seek cache. RAID seeks too much and we cacheing this. We
   fool it and just storing new position in file to _seek_vector.
   When there is no seeks to do, we are putting RAID_SEEK_DONE into it.
   Any other value requires seeking to that position.
 
   TODO:
 
 
   -  Implement other fancy things like RAID 1 (mirroring) and RAID 5.
      Should not to be very complex.
 
   -  Optimize big blob writes by resorting write buffers and writing
      big chunks at once instead of doing many syscalls. - after thinking I
      found this is useless. This is because same thing one can do with just
      increasing RAID_CHUNKSIZE. Monty, what do you think? tonu.
 
   -  If needed, then implement missing syscalls. One known to miss is stat();
 
   -  Make and use a thread safe dynamic_array buffer. The used one
      will not work if needs to be extended at the same time someone is
      accessing it.
 
 
   tonu@mysql.com & monty@mysql.com
*/

#ifdef USE_PRAGMA_IMPLEMENTATION 
#pragma implementation				// gcc: Class implementation
#endif

#include "mysys_priv.h"
#include <my_dir.h>
#include <m_string.h>
#include <assert.h>

#if defined(USE_RAID) && !defined(MYSQL_CLIENT)

#define RAID_SEEK_DONE ~(off_t) 0
#define RAID_SIZE_UNKNOWN ~(my_off_t) 0

DYNAMIC_ARRAY RaidFd::_raid_map;


/* ---------------  C compatibility  ---------------*/

extern "C" {

  void init_raid(void)
  {
  /* Allocate memory for global file to raid map */
    my_init_dynamic_array(&RaidFd::_raid_map, sizeof(RaidFd*), 4096, 1024);
  }
  void end_raid(void)
  {
    /* Free memory used by raid */
    delete_dynamic(&RaidFd::_raid_map);
  }

  bool is_raid(File fd)
  {
    return RaidFd::IsRaid(fd);
  }

  File my_raid_create(const char *FileName, int CreateFlags, int access_flags,
		      uint raid_type, uint raid_chunks, ulong raid_chunksize,
		      myf MyFlags)
  {
    DBUG_ENTER("my_raid_create");
    DBUG_PRINT("enter",("Filename: %s  CreateFlags: %d  access_flags: %d  MyFlags: %d",
			FileName, CreateFlags, access_flags, MyFlags));
    if (raid_type)
    {
      RaidFd *raid = new RaidFd(raid_type, raid_chunks , raid_chunksize);
      File res = raid->Create(FileName,CreateFlags,access_flags,MyFlags);
      if (res < 0 || set_dynamic(&RaidFd::_raid_map,(char*) &raid,res))
      {
	delete raid;
	DBUG_RETURN(-1);
      }
      DBUG_RETURN(res);
    }
    else
       DBUG_RETURN(my_create(FileName, CreateFlags, access_flags,  MyFlags));
  }

  File my_raid_open(const char *FileName, int Flags,
		    uint raid_type, uint raid_chunks, ulong raid_chunksize,
		    myf MyFlags)
  {
    DBUG_ENTER("my_raid_open");
    DBUG_PRINT("enter",("Filename: %s  Flags: %d  MyFlags: %d",
			FileName, Flags, MyFlags));
    if (raid_type)
    {
      RaidFd *raid = new RaidFd(raid_type, raid_chunks , raid_chunksize);
      File res = raid->Open(FileName,Flags,MyFlags);
      if (res < 0 || set_dynamic(&RaidFd::_raid_map,(char*) &raid,res))
      {
	delete raid;
	DBUG_RETURN(-1);
      }
      DBUG_RETURN(res);
    }
    else
      DBUG_RETURN(my_open(FileName, Flags, MyFlags));
  }

  my_off_t my_raid_seek(File fd, my_off_t pos,int whence,myf MyFlags)
  {
    DBUG_ENTER("my_raid_seek");
    DBUG_PRINT("enter",("Fd: %d  pos: %lu whence: %d  MyFlags: %d",
			fd, (ulong) pos, whence, MyFlags));

    if (is_raid(fd))
    {
      assert(pos != MY_FILEPOS_ERROR);

      RaidFd *raid= (*dynamic_element(&RaidFd::_raid_map,fd,RaidFd**));
      DBUG_RETURN(raid->Seek(pos,whence,MyFlags));
    }
    else
      DBUG_RETURN(my_seek(fd, pos, whence, MyFlags));
  }

  my_off_t my_raid_tell(File fd,myf MyFlags)
  {
    DBUG_ENTER("my_raid_tell");
    DBUG_PRINT("enter",("Fd: %d  MyFlags: %d",
			fd, MyFlags));
    if (is_raid(fd))
    {
      RaidFd *raid= (*dynamic_element(&RaidFd::_raid_map,fd,RaidFd**));
      DBUG_RETURN(raid->Tell(MyFlags));
    }
    else
       DBUG_RETURN(my_tell(fd, MyFlags));
  }

  uint my_raid_write(File fd,const byte *Buffer, uint Count, myf MyFlags)
  {
    DBUG_ENTER("my_raid_write");
    DBUG_PRINT("enter",("Fd: %d  Buffer: 0x%lx  Count: %u  MyFlags: %d",
		      fd, Buffer, Count, MyFlags));
    if (is_raid(fd))
    {
      RaidFd *raid= (*dynamic_element(&RaidFd::_raid_map,fd,RaidFd**));
      DBUG_RETURN(raid->Write(Buffer,Count,MyFlags));
    } else
      DBUG_RETURN(my_write(fd,Buffer,Count,MyFlags));
  }

  uint my_raid_read(File fd, byte *Buffer, uint Count, myf MyFlags)
  {
    DBUG_ENTER("my_raid_read");
    DBUG_PRINT("enter",("Fd: %d  Buffer: 0x%lx  Count: %u  MyFlags: %d",
		      fd, Buffer, Count, MyFlags));
    if (is_raid(fd))
    {
      RaidFd *raid= (*dynamic_element(&RaidFd::_raid_map,fd,RaidFd**));
      DBUG_RETURN(raid->Read(Buffer,Count,MyFlags));
    } else
      DBUG_RETURN(my_read(fd,Buffer,Count,MyFlags));
  }

  uint my_raid_pread(File Filedes, byte *Buffer, uint Count, my_off_t offset,
		     myf MyFlags)
  {
    DBUG_ENTER("my_raid_pread");
    DBUG_PRINT("enter",
               ("Fd: %d  Buffer: 0x%lx  Count: %u offset: %u  MyFlags: %d",
                Filedes, Buffer, Count, offset, MyFlags));
     if (is_raid(Filedes))
     {
       assert(offset != MY_FILEPOS_ERROR);

       RaidFd *raid= (*dynamic_element(&RaidFd::_raid_map,Filedes,RaidFd**));
       /* Returning value isn't important because real seek is done later. */
       raid->Seek(offset,MY_SEEK_SET,MyFlags);
       DBUG_RETURN(raid->Read(Buffer,Count,MyFlags));
     }
     else
       DBUG_RETURN(my_pread(Filedes, Buffer, Count, offset, MyFlags));
  }

  uint my_raid_pwrite(int Filedes, const byte *Buffer, uint Count,
		      my_off_t offset, myf MyFlags)
  {
    DBUG_ENTER("my_raid_pwrite");
    DBUG_PRINT("enter",
               ("Fd: %d  Buffer: 0x %lx  Count: %u offset: %u  MyFlags: %d",
                Filedes, Buffer, Count, offset, MyFlags));
     if (is_raid(Filedes))
     {
       assert(offset != MY_FILEPOS_ERROR);

       RaidFd *raid= (*dynamic_element(&RaidFd::_raid_map,Filedes,RaidFd**));
       /* Returning value isn't important because real seek is done later. */
       raid->Seek(offset,MY_SEEK_SET,MyFlags);
       DBUG_RETURN(raid->Write(Buffer,Count,MyFlags));
     }
     else
       DBUG_RETURN(my_pwrite(Filedes, Buffer, Count, offset, MyFlags));
  }

  int my_raid_lock(File fd, int locktype, my_off_t start, my_off_t length,
		   myf MyFlags)
  {
    DBUG_ENTER("my_raid_lock");
    DBUG_PRINT("enter",("Fd: %d  start: %u  length: %u  MyFlags: %d",
		      fd, start, length, MyFlags));
    if (my_disable_locking)
      DBUG_RETURN(0);
    if (is_raid(fd))
    {
      RaidFd *raid= (*dynamic_element(&RaidFd::_raid_map,fd,RaidFd**));
      DBUG_RETURN(raid->Lock(locktype, start, length, MyFlags));
    }
    else
      DBUG_RETURN(my_lock(fd, locktype, start, length, MyFlags));
  }

  int my_raid_close(File fd, myf MyFlags)
  {
    DBUG_ENTER("my_raid_close");
    DBUG_PRINT("enter",("Fd: %d  MyFlags: %d",
		      fd, MyFlags));
    if (is_raid(fd))
    {
      RaidFd *raid= (*dynamic_element(&RaidFd::_raid_map,fd,RaidFd**));
      RaidFd *tmp=0;
      set_dynamic(&RaidFd::_raid_map,(char*) &tmp,fd);
      int res = raid->Close(MyFlags);
      delete raid;
      DBUG_RETURN(res);
    }
    else
      DBUG_RETURN(my_close(fd, MyFlags));
  }

  int my_raid_chsize(File fd, my_off_t newlength, int filler, myf MyFlags)
  {
    DBUG_ENTER("my_raid_chsize");
    DBUG_PRINT("enter",("Fd: %d  newlength: %u  MyFlags: %d",
		      fd, newlength, MyFlags));
   if (is_raid(fd))
   {
     RaidFd *raid= (*dynamic_element(&RaidFd::_raid_map,fd,RaidFd**));
     DBUG_RETURN(raid->Chsize(fd, newlength, filler, MyFlags));
   }
   else
     DBUG_RETURN(my_chsize(fd, newlength, filler, MyFlags));
  }

  int my_raid_rename(const char *from, const char *to,
		     uint raid_chunks, myf MyFlags)
  {
    char from_tmp[FN_REFLEN];
    char to_tmp[FN_REFLEN];
    DBUG_ENTER("my_raid_rename");

    uint from_pos = dirname_length(from);
    uint to_pos   = dirname_length(to);
    memcpy(from_tmp, from, from_pos);
    memcpy(to_tmp, to, to_pos);
    for (uint i = 0 ; i < raid_chunks ; i++ )
    {
      sprintf(from_tmp+from_pos,"%02x/%s", i, from + from_pos);
      sprintf(to_tmp+to_pos,"%02x/%s", i, to+ to_pos);
      /* Convert if not unix */
      unpack_filename(from_tmp, from_tmp);
      unpack_filename(to_tmp,to_tmp);
      if (my_rename(from_tmp, to_tmp, MyFlags))
	DBUG_RETURN(-1);
    }
    DBUG_RETURN(0);
  }

  int my_raid_delete(const char *from, uint raid_chunks, myf MyFlags)
  {
    char from_tmp[FN_REFLEN];
    uint from_pos = dirname_length(from);
    DBUG_ENTER("my_raid_delete");

    if (!raid_chunks)
      DBUG_RETURN(my_delete(from,MyFlags));
    for (uint i = 0 ; i < raid_chunks ; i++ )
    {
      memcpy(from_tmp, from, from_pos);
      sprintf(from_tmp+from_pos,"%02x/%s", i, from + from_pos);
      /* Convert if not unix */
      unpack_filename(from_tmp, from_tmp);
      if (my_delete(from_tmp, MyFlags))
	DBUG_RETURN(-1);
    }
    DBUG_RETURN(0);
  }

  int my_raid_redel(const char *old_name, const char *new_name,
		    uint raid_chunks, myf MyFlags)
  {
    char new_name_buff[FN_REFLEN], old_name_buff[FN_REFLEN];
    char *new_end, *old_end;
    uint i,old_length,new_length;
    int error=0;
    DBUG_ENTER("my_raid_redel");

    old_end=old_name_buff+dirname_part(old_name_buff,old_name);
    old_length=dirname_length(old_name);
    new_end=new_name_buff+dirname_part(new_name_buff,new_name);
    new_length=dirname_length(new_name);
    for (i=0 ;	i < raid_chunks ; i++)
    {
      MY_STAT status;
      sprintf(new_end,"%02x",i);
      if (my_stat(new_name_buff,&status, MYF(0)))
      {
	DBUG_PRINT("info",("%02x exists, skipping directory creation",i));
      }
      else
      {
	if (my_mkdir(new_name_buff,0777,MYF(0)))
	{
	  DBUG_PRINT("error",("mkdir failed for %02x",i));
	  DBUG_RETURN(-1);
	}
      }
      strxmov(strend(new_end),"/",new_name+new_length,NullS);
      sprintf(old_end,"%02x/%s",i, old_name+old_length);
      if (my_redel(old_name_buff, new_name_buff, MyFlags))
	error=1;
    }
    DBUG_RETURN(error);
  }
}

int my_raid_fstat(int fd, MY_STAT *stat_area, myf MyFlags )
{
  DBUG_ENTER("my_raid_fstat");
  if (is_raid(fd))
  {
    RaidFd *raid= (*dynamic_element(&RaidFd::_raid_map,fd,RaidFd**));
    DBUG_RETURN(raid->Fstat(fd, stat_area, MyFlags));
  }
  else
    DBUG_RETURN(my_fstat(fd, stat_area, MyFlags));
}


/* -------------- RaidFd base class begins ----------------*/
/*
  RaidFd - raided file is identified by file descriptor
  this is useful when we open/write/read/close files
*/


bool RaidFd::
IsRaid(File fd)
{
  DBUG_ENTER("RaidFd::IsRaid");
  DBUG_RETURN((uint) fd < _raid_map.elements &&
	      *dynamic_element(&RaidFd::_raid_map,fd,RaidFd**));
}


RaidFd::
RaidFd(uint raid_type, uint raid_chunks, ulong raid_chunksize)
  :_raid_type(raid_type), _raid_chunks(raid_chunks),
   _raid_chunksize(raid_chunksize), _position(0), _size(RAID_SIZE_UNKNOWN),
   _fd_vector(0)
{
  DBUG_ENTER("RaidFd::RaidFd");
  DBUG_PRINT("enter",("RaidFd_type: %u  Disks: %u  Chunksize: %d",
		   raid_type, raid_chunks, raid_chunksize));

  /* TODO: Here we should add checks if the malloc fails */
  _seek_vector=0;				/* In case of errors */
  my_multi_malloc(MYF(MY_WME),
		  &_seek_vector,sizeof(off_t)*_raid_chunks,
		  &_fd_vector, sizeof(File) *_raid_chunks,
		  NullS);
  if (!RaidFd::_raid_map.buffer)
  {					/* Not initied */
    pthread_mutex_lock(&THR_LOCK_open);	/* Ensure that no other thread */
    if (!RaidFd::_raid_map.buffer)	/* has done init in between */
      init_raid();
    pthread_mutex_unlock(&THR_LOCK_open);
  }
  DBUG_VOID_RETURN;
}


RaidFd::
~RaidFd() {
  DBUG_ENTER("RaidFd::~RaidFd");
  /* We don't have to free _fd_vector ! */
  my_free((char*) _seek_vector, MYF(MY_ALLOW_ZERO_PTR));
  DBUG_VOID_RETURN;
}


File RaidFd::
Create(const char *FileName, int CreateFlags, int access_flags, myf MyFlags)
{
  char RaidFdFileName[FN_REFLEN];
  DBUG_ENTER("RaidFd::Create");
  DBUG_PRINT("enter",
	     ("FileName: %s  CreateFlags: %d  access_flags: %d  MyFlags: %d",
	      FileName, CreateFlags, access_flags, MyFlags));
  char DirName[FN_REFLEN];
  uint pos = dirname_part(DirName, FileName);
  MY_STAT status;
  if (!_seek_vector)
    DBUG_RETURN(-1);				/* Not enough memory */

  uint i = _raid_chunks-1;
  do
  {
    /* Create subdir */
    (void)sprintf(RaidFdFileName,"%s%02x", DirName,i);
    unpack_dirname(RaidFdFileName,RaidFdFileName);   /* Convert if not unix */
    if (my_stat(RaidFdFileName,&status, MYF(0)))
    {
      DBUG_PRINT("info",("%02x exists, skipping directory creation",i));
    }
    else
    {
      if (my_mkdir(RaidFdFileName,0777,MYF(0)))
      {
	DBUG_PRINT("error",("mkdir failed for %d",i));
	goto error;
      }
    }
    /* Create file */
    sprintf(RaidFdFileName,"%s%02x/%s",DirName, i, FileName + pos);
    unpack_filename(RaidFdFileName,RaidFdFileName); /* Convert if not unix */
    _fd = my_create(RaidFdFileName, CreateFlags ,access_flags, (myf)MyFlags);
    if (_fd < 0)
      goto error;
    _fd_vector[i]=_fd;
    _seek_vector[i]=RAID_SEEK_DONE;
  } while (i--);
  _size=0;
  DBUG_RETURN(_fd);			/* Last filenr is pointer to map */

error:
  {
    int save_errno=my_errno;
    while (++i < _raid_chunks)
    {
      my_close(_fd_vector[i],MYF(0));
      sprintf(RaidFdFileName,"%s%02x/%s",DirName, i, FileName + pos);
      unpack_filename(RaidFdFileName,RaidFdFileName);
      my_delete(RaidFdFileName,MYF(0));
    }
    my_errno=save_errno;
  }
  DBUG_RETURN(-1);
}


File RaidFd::
Open(const char *FileName, int Flags, myf MyFlags)
{
  DBUG_ENTER("RaidFd::Open");
  DBUG_PRINT("enter",("FileName: %s  Flags: %d  MyFlags: %d",
		   FileName, Flags, MyFlags));
  char DirName[FN_REFLEN];
  uint pos = dirname_part(DirName, FileName);
  if (!_seek_vector)
    DBUG_RETURN(-1);				/* Not enough memory */

  for( uint i = 0 ;  i < _raid_chunks ; i++ )
  {
    char RaidFdFileName[FN_REFLEN];
    sprintf(RaidFdFileName,"%s%02x/%s",DirName, i, FileName + pos);
    unpack_filename(RaidFdFileName,RaidFdFileName); /* Convert if not unix */
    _fd = my_open(RaidFdFileName, Flags, MyFlags);
    if (_fd < 0)
    {
      int save_errno=my_errno;
      while (i-- != 0)
	my_close(_fd_vector[i],MYF(0));
      my_errno=save_errno;
      DBUG_RETURN(_fd);
    }
    _fd_vector[i]=_fd;
    _seek_vector[i]=RAID_SEEK_DONE;
  }
  Seek(0L,MY_SEEK_END,MYF(0)); // Trick. We just need to know, how big the file is
  DBUG_PRINT("info",("MYD file logical size: %llu", _size));
  DBUG_RETURN(_fd);
}


int RaidFd::
Write(const byte *Buffer, uint Count, myf MyFlags)
{
  DBUG_ENTER("RaidFd::Write");
  DBUG_PRINT("enter",("Count: %d  MyFlags: %d",
		      Count, MyFlags));
  const byte *bufptr = Buffer;
  uint res=0, GotBytes, ReadNowCount;

  // Loop until data is written
  do {
    Calculate();
     // Do seeks when neccessary
    if (_seek_vector[_this_block] != RAID_SEEK_DONE)
    {
      if (my_seek(_fd_vector[_this_block], _seek_vector[_this_block],
		  MY_SEEK_SET,
		  MyFlags) == MY_FILEPOS_ERROR)
	DBUG_RETURN(-1);
      _seek_vector[_this_block]=RAID_SEEK_DONE;
    }
    ReadNowCount = min(Count, _remaining_bytes);
    GotBytes = my_write(_fd_vector[_this_block], bufptr, ReadNowCount,
			MyFlags);
    DBUG_PRINT("loop",("Wrote bytes: %d", GotBytes));
    if (GotBytes == MY_FILE_ERROR)
      DBUG_RETURN(-1);
    res+= GotBytes;
    if (MyFlags & (MY_NABP | MY_FNABP))
      GotBytes=ReadNowCount;
    bufptr += GotBytes;
    Count  -= GotBytes;
    _position += GotBytes;
  } while(Count);
  set_if_bigger(_size,_position);
  DBUG_RETURN(res);
}


int RaidFd::
Read(const byte *Buffer, uint Count, myf MyFlags)
{
  DBUG_ENTER("RaidFd::Read");
  DBUG_PRINT("enter",("Count: %d  MyFlags: %d",
		      Count, MyFlags));
  byte *bufptr = (byte *)Buffer;
  uint res= 0, GotBytes, ReadNowCount;

  // Loop until all data is read (Note that Count may be 0)
  while (Count)
  {
    Calculate();
    // Do seek when neccessary
    if (_seek_vector[_this_block] != RAID_SEEK_DONE)
    {
      if (my_seek(_fd_vector[_this_block], _seek_vector[_this_block],
		  MY_SEEK_SET,
		  MyFlags) == MY_FILEPOS_ERROR)
	DBUG_RETURN(-1);
      _seek_vector[_this_block]=RAID_SEEK_DONE;
    }
    // and read
    ReadNowCount = min(Count, _remaining_bytes);
    GotBytes = my_read(_fd_vector[_this_block], bufptr, ReadNowCount,
		       MyFlags & ~(MY_NABP | MY_FNABP));
    DBUG_PRINT("loop",("Got bytes: %u", GotBytes));
    if (GotBytes == MY_FILE_ERROR)
      DBUG_RETURN(-1);
    if (!GotBytes)				// End of file.
    {
      DBUG_RETURN((MyFlags & (MY_NABP | MY_FNABP)) ? -1 : (int) res);
    }
    res+= GotBytes;
    bufptr += GotBytes;
    Count  -= GotBytes;
    _position += GotBytes;
  }
  DBUG_RETURN((MyFlags & (MY_NABP | MY_FNABP)) ? 0 : res);
}


int RaidFd::
Lock(int locktype, my_off_t start, my_off_t length, myf MyFlags)
{
  DBUG_ENTER("RaidFd::Lock");
  DBUG_PRINT("enter",("locktype: %d  start: %lu  length: %lu  MyFlags: %d",
		      locktype, start, length, MyFlags));
  my_off_t bufptr = start;
  // Loop until all data is locked
  while(length)
  {
    Calculate();
    for (uint i = _this_block ; (i < _raid_chunks) && length ; i++ )
    {
       uint ReadNowCount = min(length, _remaining_bytes);
       uint GotBytes = my_lock(_fd_vector[i], locktype, bufptr, ReadNowCount,
			MyFlags);
       if ((int) GotBytes == -1)
	 DBUG_RETURN(-1);
       bufptr += ReadNowCount;
       length  -= ReadNowCount;
       Calculate();
    }
  }
  DBUG_RETURN(0);
}


int RaidFd::
Close(myf MyFlags)
{
  DBUG_ENTER("RaidFd::Close");
  DBUG_PRINT("enter",("MyFlags: %d",
		   MyFlags));
  for (uint i = 0 ; i < _raid_chunks ; ++i )
  {
    int err = my_close(_fd_vector[i], MyFlags);
    if (err != 0)
      DBUG_RETURN(err);
  }
  /* _fd_vector is erased when RaidFd is released */
  DBUG_RETURN(0);
}


my_off_t RaidFd::
Seek(my_off_t pos,int whence,myf MyFlags)
{
  DBUG_ENTER("RaidFd::Seek");
  DBUG_PRINT("enter",("Pos: %lu  Whence: %d  MyFlags: %d",
		   (ulong) pos, whence, MyFlags));
  switch (whence) {
  case MY_SEEK_CUR:
    // FIXME: This is wrong, what is going on there
    // Just I am relied on fact that MySQL 3.23.7 never uses MY_SEEK_CUR
    // for anything else except things like ltell()
    break;
  case MY_SEEK_SET:
    if ( _position != pos) // we can be already in right place
    {
      uint i;
      off_t _rounds;
      _position = pos;
      Calculate();
      _rounds = _total_block / _raid_chunks;	    // INT() assumed
      _rounds*= _raid_chunksize;
      for (i = 0; i < _raid_chunks ; i++ )
	if ( i < _this_block )
	  _seek_vector[i] = _rounds + _raid_chunksize;
	else if ( i == _this_block )
	  _seek_vector[i] = _rounds + _raid_chunksize -_remaining_bytes;
	else					// if ( i > _this_block )
	  _seek_vector[i] = _rounds;
    }
    break;
  case MY_SEEK_END:
    if (_size==RAID_SIZE_UNKNOWN) // We don't know table size yet
    {
      uint i;
      _position = 0;
      for (i = 0; i < _raid_chunks ; i++ )
      {
	my_off_t newpos = my_seek(_fd_vector[i], 0L, MY_SEEK_END, MyFlags);
	if (newpos == MY_FILEPOS_ERROR)
	  DBUG_RETURN (MY_FILEPOS_ERROR);
	_seek_vector[i]=RAID_SEEK_DONE;
	_position += newpos;
      }
      _size=_position;
    }
    else if (_position != _size) // Aren't we also already in the end?
    {
      uint i;
      off_t _rounds;
      _position = _size;
      Calculate();
      _rounds = _total_block / _raid_chunks;	    // INT() assumed
      _rounds*= _raid_chunksize;
      for (i = 0; i < _raid_chunks ; i++ )
	if ( i < _this_block )
	  _seek_vector[i] = _rounds + _raid_chunksize;
	else if ( i == _this_block )
	  _seek_vector[i] = _rounds + _raid_chunksize - _remaining_bytes;
	else					// if ( i > _this_block )
	  _seek_vector[i] = _rounds;
      _position=_size;
    }
  }
  DBUG_RETURN(_position);
}


my_off_t RaidFd::
Tell(myf MyFlags)
{
  DBUG_ENTER("RaidFd::Tell");
  DBUG_PRINT("enter",("MyFlags: %d _position %d",
		   MyFlags,_position));
  DBUG_RETURN(_position);
}

int RaidFd::
Chsize(File fd, my_off_t newlength, int filler, myf MyFlags)
{
  DBUG_ENTER("RaidFd::Chsize");
  DBUG_PRINT("enter",("Fd: %d, newlength: %d, MyFlags: %d",
		   fd, newlength,MyFlags));
  _position = newlength;
  Calculate();
  uint _rounds = _total_block / _raid_chunks;	     // INT() assumed
  for (uint i = 0; i < _raid_chunks ; i++ )
  {
    int newpos;
    if ( i < _this_block )
      newpos = my_chsize(_fd_vector[i],
			 _this_block * _raid_chunksize + (_rounds + 1) *
			 _raid_chunksize, filler, MyFlags);
    else if ( i == _this_block )
      newpos = my_chsize(_fd_vector[i],
			 _this_block * _raid_chunksize + _rounds *
			 _raid_chunksize + (newlength % _raid_chunksize),
			 filler, MyFlags);
    else // this means: i > _this_block
      newpos = my_chsize(_fd_vector[i],
			 _this_block * _raid_chunksize + _rounds *
			 _raid_chunksize, filler, MyFlags);
    if (newpos)
      DBUG_RETURN(1);
  }
  DBUG_RETURN(0);
}


int RaidFd::
Fstat(int fd, MY_STAT *stat_area, myf MyFlags )
{
  DBUG_ENTER("RaidFd::Fstat");
  DBUG_PRINT("enter",("fd: %d MyFlags: %d",fd,MyFlags));
  uint i;
  int error=0;
  MY_STAT status;
  stat_area->st_size=0;
  stat_area->st_mtime=0;
  stat_area->st_atime=0;
  stat_area->st_ctime=0;

  for(i=0 ; i < _raid_chunks ; i++)
  {
    if (my_fstat(_fd_vector[i],&status,MyFlags))
      error=1;
    stat_area->st_size+=status.st_size;
    set_if_bigger(stat_area->st_mtime,status.st_mtime);
    set_if_bigger(stat_area->st_atime,status.st_atime);
    set_if_bigger(stat_area->st_ctime,status.st_ctime);
  }
  DBUG_RETURN(error);
}

#endif /* defined(USE_RAID) && !defined(MYSQL_CLIENT) */