Training courses

Kernel and Embedded Linux

Bootlin training courses

Embedded Linux, kernel,
Yocto Project, Buildroot, real-time,
graphics, boot time, debugging...

Bootlin logo

Elixir Cross Referencer

  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
/*	$NetBSD: rf_dagfuncs.c,v 1.30 2009/03/23 18:38:54 oster Exp $	*/
/*
 * Copyright (c) 1995 Carnegie-Mellon University.
 * All rights reserved.
 *
 * Author: Mark Holland, William V. Courtright II
 *
 * Permission to use, copy, modify and distribute this software and
 * its documentation is hereby granted, provided that both the copyright
 * notice and this permission notice appear in all copies of the
 * software, derivative works or modified versions, and any portions
 * thereof, and that both notices appear in supporting documentation.
 *
 * CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
 * CONDITION.  CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND
 * FOR ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
 *
 * Carnegie Mellon requests users of this software to return to
 *
 *  Software Distribution Coordinator  or  Software.Distribution@CS.CMU.EDU
 *  School of Computer Science
 *  Carnegie Mellon University
 *  Pittsburgh PA 15213-3890
 *
 * any improvements or extensions that they make and grant Carnegie the
 * rights to redistribute these changes.
 */

/*
 * dagfuncs.c -- DAG node execution routines
 *
 * Rules:
 * 1. Every DAG execution function must eventually cause node->status to
 *    get set to "good" or "bad", and "FinishNode" to be called. In the
 *    case of nodes that complete immediately (xor, NullNodeFunc, etc),
 *    the node execution function can do these two things directly. In
 *    the case of nodes that have to wait for some event (a disk read to
 *    complete, a lock to be released, etc) to occur before they can
 *    complete, this is typically achieved by having whatever module
 *    is doing the operation call GenericWakeupFunc upon completion.
 * 2. DAG execution functions should check the status in the DAG header
 *    and NOP out their operations if the status is not "enable". However,
 *    execution functions that release resources must be sure to release
 *    them even when they NOP out the function that would use them.
 *    Functions that acquire resources should go ahead and acquire them
 *    even when they NOP, so that a downstream release node will not have
 *    to check to find out whether or not the acquire was suppressed.
 */

#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: rf_dagfuncs.c,v 1.30 2009/03/23 18:38:54 oster Exp $");

#include <sys/param.h>
#include <sys/ioctl.h>

#include "rf_archs.h"
#include "rf_raid.h"
#include "rf_dag.h"
#include "rf_layout.h"
#include "rf_etimer.h"
#include "rf_acctrace.h"
#include "rf_diskqueue.h"
#include "rf_dagfuncs.h"
#include "rf_general.h"
#include "rf_engine.h"
#include "rf_dagutils.h"

#include "rf_kintf.h"

#if RF_INCLUDE_PARITYLOGGING > 0
#include "rf_paritylog.h"
#endif				/* RF_INCLUDE_PARITYLOGGING > 0 */

int     (*rf_DiskReadFunc) (RF_DagNode_t *);
int     (*rf_DiskWriteFunc) (RF_DagNode_t *);
int     (*rf_DiskReadUndoFunc) (RF_DagNode_t *);
int     (*rf_DiskWriteUndoFunc) (RF_DagNode_t *);
int     (*rf_RegularXorUndoFunc) (RF_DagNode_t *);
int     (*rf_SimpleXorUndoFunc) (RF_DagNode_t *);
int     (*rf_RecoveryXorUndoFunc) (RF_DagNode_t *);

/*****************************************************************************
 * main (only) configuration routine for this module
 ****************************************************************************/
int
rf_ConfigureDAGFuncs(RF_ShutdownList_t **listp)
{
	RF_ASSERT(((sizeof(long) == 8) && RF_LONGSHIFT == 3) ||
		  ((sizeof(long) == 4) && RF_LONGSHIFT == 2));
	rf_DiskReadFunc = rf_DiskReadFuncForThreads;
	rf_DiskReadUndoFunc = rf_DiskUndoFunc;
	rf_DiskWriteFunc = rf_DiskWriteFuncForThreads;
	rf_DiskWriteUndoFunc = rf_DiskUndoFunc;
	rf_RegularXorUndoFunc = rf_NullNodeUndoFunc;
	rf_SimpleXorUndoFunc = rf_NullNodeUndoFunc;
	rf_RecoveryXorUndoFunc = rf_NullNodeUndoFunc;
	return (0);
}



/*****************************************************************************
 * the execution function associated with a terminate node
 ****************************************************************************/
int
rf_TerminateFunc(RF_DagNode_t *node)
{
	RF_ASSERT(node->dagHdr->numCommits == node->dagHdr->numCommitNodes);
	node->status = rf_good;
	return (rf_FinishNode(node, RF_THREAD_CONTEXT));
}

int
rf_TerminateUndoFunc(RF_DagNode_t *node)
{
	return (0);
}


/*****************************************************************************
 * execution functions associated with a mirror node
 *
 * parameters:
 *
 * 0 - physical disk addres of data
 * 1 - buffer for holding read data
 * 2 - parity stripe ID
 * 3 - flags
 * 4 - physical disk address of mirror (parity)
 *
 ****************************************************************************/

int
rf_DiskReadMirrorIdleFunc(RF_DagNode_t *node)
{
	/* select the mirror copy with the shortest queue and fill in node
	 * parameters with physical disk address */

	rf_SelectMirrorDiskIdle(node);
	return (rf_DiskReadFunc(node));
}

#if (RF_INCLUDE_CHAINDECLUSTER > 0) || (RF_INCLUDE_INTERDECLUSTER > 0) || (RF_DEBUG_VALIDATE_DAG > 0)
int
rf_DiskReadMirrorPartitionFunc(RF_DagNode_t *node)
{
	/* select the mirror copy with the shortest queue and fill in node
	 * parameters with physical disk address */

	rf_SelectMirrorDiskPartition(node);
	return (rf_DiskReadFunc(node));
}
#endif

int
rf_DiskReadMirrorUndoFunc(RF_DagNode_t *node)
{
	return (0);
}



#if RF_INCLUDE_PARITYLOGGING > 0
/*****************************************************************************
 * the execution function associated with a parity log update node
 ****************************************************************************/
int
rf_ParityLogUpdateFunc(RF_DagNode_t *node)
{
	RF_PhysDiskAddr_t *pda = (RF_PhysDiskAddr_t *) node->params[0].p;
	void *bf = (void *) node->params[1].p;
	RF_ParityLogData_t *logData;
#if RF_ACC_TRACE > 0
	RF_AccTraceEntry_t *tracerec = node->dagHdr->tracerec;
	RF_Etimer_t timer;
#endif

	if (node->dagHdr->status == rf_enable) {
#if RF_ACC_TRACE > 0
		RF_ETIMER_START(timer);
#endif
		logData = rf_CreateParityLogData(RF_UPDATE, pda, bf,
		    (RF_Raid_t *) (node->dagHdr->raidPtr),
		    node->wakeFunc, (void *) node,
		    node->dagHdr->tracerec, timer);
		if (logData)
			rf_ParityLogAppend(logData, RF_FALSE, NULL, RF_FALSE);
		else {
#if RF_ACC_TRACE > 0
			RF_ETIMER_STOP(timer);
			RF_ETIMER_EVAL(timer);
			tracerec->plog_us += RF_ETIMER_VAL_US(timer);
#endif
			(node->wakeFunc) (node, ENOMEM);
		}
	}
	return (0);
}


/*****************************************************************************
 * the execution function associated with a parity log overwrite node
 ****************************************************************************/
int
rf_ParityLogOverwriteFunc(RF_DagNode_t *node)
{
	RF_PhysDiskAddr_t *pda = (RF_PhysDiskAddr_t *) node->params[0].p;
	void *bf = (void *) node->params[1].p;
	RF_ParityLogData_t *logData;
#if RF_ACC_TRACE > 0
	RF_AccTraceEntry_t *tracerec = node->dagHdr->tracerec;
	RF_Etimer_t timer;
#endif

	if (node->dagHdr->status == rf_enable) {
#if RF_ACC_TRACE > 0
		RF_ETIMER_START(timer);
#endif
		logData = rf_CreateParityLogData(RF_OVERWRITE, pda, bf,
(RF_Raid_t *) (node->dagHdr->raidPtr),
		    node->wakeFunc, (void *) node, node->dagHdr->tracerec, timer);
		if (logData)
			rf_ParityLogAppend(logData, RF_FALSE, NULL, RF_FALSE);
		else {
#if RF_ACC_TRACE > 0
			RF_ETIMER_STOP(timer);
			RF_ETIMER_EVAL(timer);
			tracerec->plog_us += RF_ETIMER_VAL_US(timer);
#endif
			(node->wakeFunc) (node, ENOMEM);
		}
	}
	return (0);
}

int
rf_ParityLogUpdateUndoFunc(RF_DagNode_t *node)
{
	return (0);
}

int
rf_ParityLogOverwriteUndoFunc(RF_DagNode_t *node)
{
	return (0);
}
#endif				/* RF_INCLUDE_PARITYLOGGING > 0 */

/*****************************************************************************
 * the execution function associated with a NOP node
 ****************************************************************************/
int
rf_NullNodeFunc(RF_DagNode_t *node)
{
	node->status = rf_good;
	return (rf_FinishNode(node, RF_THREAD_CONTEXT));
}

int
rf_NullNodeUndoFunc(RF_DagNode_t *node)
{
	node->status = rf_undone;
	return (rf_FinishNode(node, RF_THREAD_CONTEXT));
}


/*****************************************************************************
 * the execution function associated with a disk-read node
 ****************************************************************************/
int
rf_DiskReadFuncForThreads(RF_DagNode_t *node)
{
	RF_DiskQueueData_t *req;
	RF_PhysDiskAddr_t *pda = (RF_PhysDiskAddr_t *) node->params[0].p;
	void *bf = (void *) node->params[1].p;
	RF_StripeNum_t parityStripeID = (RF_StripeNum_t) node->params[2].v;
	unsigned priority = RF_EXTRACT_PRIORITY(node->params[3].v);
	unsigned which_ru = RF_EXTRACT_RU(node->params[3].v);
	RF_IoType_t iotype = (node->dagHdr->status == rf_enable) ? RF_IO_TYPE_READ : RF_IO_TYPE_NOP;
	RF_DiskQueue_t *dqs = ((RF_Raid_t *) (node->dagHdr->raidPtr))->Queues;
	void   *b_proc = NULL;

	if (node->dagHdr->bp)
		b_proc = (void *) ((struct buf *) node->dagHdr->bp)->b_proc;

	req = rf_CreateDiskQueueData(iotype, pda->startSector, pda->numSector,
	    bf, parityStripeID, which_ru,
	    (int (*) (void *, int)) node->wakeFunc,
	    node,
#if RF_ACC_TRACE > 0
	     node->dagHdr->tracerec,
#else
             NULL,
#endif
	    (void *) (node->dagHdr->raidPtr), 0, b_proc, PR_NOWAIT);
	if (!req) {
		(node->wakeFunc) (node, ENOMEM);
	} else {
		node->dagFuncData = (void *) req;
		rf_DiskIOEnqueue(&(dqs[pda->col]), req, priority);
	}
	return (0);
}


/*****************************************************************************
 * the execution function associated with a disk-write node
 ****************************************************************************/
int
rf_DiskWriteFuncForThreads(RF_DagNode_t *node)
{
	RF_DiskQueueData_t *req;
	RF_PhysDiskAddr_t *pda = (RF_PhysDiskAddr_t *) node->params[0].p;
	void *bf = (void *) node->params[1].p;
	RF_StripeNum_t parityStripeID = (RF_StripeNum_t) node->params[2].v;
	unsigned priority = RF_EXTRACT_PRIORITY(node->params[3].v);
	unsigned which_ru = RF_EXTRACT_RU(node->params[3].v);
	RF_IoType_t iotype = (node->dagHdr->status == rf_enable) ? RF_IO_TYPE_WRITE : RF_IO_TYPE_NOP;
	RF_DiskQueue_t *dqs = ((RF_Raid_t *) (node->dagHdr->raidPtr))->Queues;
	void   *b_proc = NULL;

	if (node->dagHdr->bp)
		b_proc = (void *) ((struct buf *) node->dagHdr->bp)->b_proc;

	/* normal processing (rollaway or forward recovery) begins here */
	req = rf_CreateDiskQueueData(iotype, pda->startSector, pda->numSector,
	    bf, parityStripeID, which_ru,
	    (int (*) (void *, int)) node->wakeFunc,
	    (void *) node,
#if RF_ACC_TRACE > 0
	    node->dagHdr->tracerec,
#else
	    NULL,
#endif
	    (void *) (node->dagHdr->raidPtr),
	    0, b_proc, PR_NOWAIT);

	if (!req) {
		(node->wakeFunc) (node, ENOMEM);
	} else {
		node->dagFuncData = (void *) req;
		rf_DiskIOEnqueue(&(dqs[pda->col]), req, priority);
	}

	return (0);
}
/*****************************************************************************
 * the undo function for disk nodes
 * Note:  this is not a proper undo of a write node, only locks are released.
 *        old data is not restored to disk!
 ****************************************************************************/
int
rf_DiskUndoFunc(RF_DagNode_t *node)
{
	RF_DiskQueueData_t *req;
	RF_PhysDiskAddr_t *pda = (RF_PhysDiskAddr_t *) node->params[0].p;
	RF_DiskQueue_t *dqs = ((RF_Raid_t *) (node->dagHdr->raidPtr))->Queues;

	req = rf_CreateDiskQueueData(RF_IO_TYPE_NOP,
	    0L, 0, NULL, 0L, 0,
	    (int (*) (void *, int)) node->wakeFunc,
	    (void *) node,
#if RF_ACC_TRACE > 0
	     node->dagHdr->tracerec,
#else
	     NULL,
#endif
	    (void *) (node->dagHdr->raidPtr),
	    0, NULL, PR_NOWAIT);
	if (!req)
		(node->wakeFunc) (node, ENOMEM);
	else {
		node->dagFuncData = (void *) req;
		rf_DiskIOEnqueue(&(dqs[pda->col]), req, RF_IO_NORMAL_PRIORITY);
	}

	return (0);
}

/*****************************************************************************
 * Callback routine for DiskRead and DiskWrite nodes.  When the disk
 * op completes, the routine is called to set the node status and
 * inform the execution engine that the node has fired.
 ****************************************************************************/
int
rf_GenericWakeupFunc(RF_DagNode_t *node, int status)
{

	switch (node->status) {
	case rf_fired:
		if (status)
			node->status = rf_bad;
		else
			node->status = rf_good;
		break;
	case rf_recover:
		/* probably should never reach this case */
		if (status)
			node->status = rf_panic;
		else
			node->status = rf_undone;
		break;
	default:
		printf("rf_GenericWakeupFunc:");
		printf("node->status is %d,", node->status);
		printf("status is %d \n", status);
		RF_PANIC();
		break;
	}
	if (node->dagFuncData)
		rf_FreeDiskQueueData((RF_DiskQueueData_t *) node->dagFuncData);
	return (rf_FinishNode(node, RF_INTR_CONTEXT));
}


/*****************************************************************************
 * there are three distinct types of xor nodes:

 * A "regular xor" is used in the fault-free case where the access
 * spans a complete stripe unit.  It assumes that the result buffer is
 * one full stripe unit in size, and uses the stripe-unit-offset
 * values that it computes from the PDAs to determine where within the
 * stripe unit to XOR each argument buffer.
 *
 * A "simple xor" is used in the fault-free case where the access
 * touches only a portion of one (or two, in some cases) stripe
 * unit(s).  It assumes that all the argument buffers are of the same
 * size and have the same stripe unit offset.
 *
 * A "recovery xor" is used in the degraded-mode case.  It's similar
 * to the regular xor function except that it takes the failed PDA as
 * an additional parameter, and uses it to determine what portions of
 * the argument buffers need to be xor'd into the result buffer, and
 * where in the result buffer they should go.
 ****************************************************************************/

/* xor the params together and store the result in the result field.
 * assume the result field points to a buffer that is the size of one
 * SU, and use the pda params to determine where within the buffer to
 * XOR the input buffers.  */
int
rf_RegularXorFunc(RF_DagNode_t *node)
{
	RF_Raid_t *raidPtr = (RF_Raid_t *) node->params[node->numParams - 1].p;
#if RF_ACC_TRACE > 0
	RF_AccTraceEntry_t *tracerec = node->dagHdr->tracerec;
	RF_Etimer_t timer;
#endif
	int     i, retcode;

	retcode = 0;
	if (node->dagHdr->status == rf_enable) {
		/* don't do the XOR if the input is the same as the output */
#if RF_ACC_TRACE > 0
		RF_ETIMER_START(timer);
#endif
		for (i = 0; i < node->numParams - 1; i += 2)
			if (node->params[i + 1].p != node->results[0]) {
				retcode = rf_XorIntoBuffer(raidPtr, (RF_PhysDiskAddr_t *) node->params[i].p,
							   (char *) node->params[i + 1].p, (char *) node->results[0]);
			}
#if RF_ACC_TRACE > 0
		RF_ETIMER_STOP(timer);
		RF_ETIMER_EVAL(timer);
		tracerec->xor_us += RF_ETIMER_VAL_US(timer);
#endif
	}
	return (rf_GenericWakeupFunc(node, retcode));	/* call wake func
							 * explicitly since no
							 * I/O in this node */
}
/* xor the inputs into the result buffer, ignoring placement issues */
int
rf_SimpleXorFunc(RF_DagNode_t *node)
{
	RF_Raid_t *raidPtr = (RF_Raid_t *) node->params[node->numParams - 1].p;
	int     i, retcode = 0;
#if RF_ACC_TRACE > 0
	RF_AccTraceEntry_t *tracerec = node->dagHdr->tracerec;
	RF_Etimer_t timer;
#endif

	if (node->dagHdr->status == rf_enable) {
#if RF_ACC_TRACE > 0
		RF_ETIMER_START(timer);
#endif
		/* don't do the XOR if the input is the same as the output */
		for (i = 0; i < node->numParams - 1; i += 2)
			if (node->params[i + 1].p != node->results[0]) {
				retcode = rf_bxor((char *) node->params[i + 1].p, (char *) node->results[0],
				    rf_RaidAddressToByte(raidPtr, ((RF_PhysDiskAddr_t *) node->params[i].p)->numSector));
			}
#if RF_ACC_TRACE > 0
		RF_ETIMER_STOP(timer);
		RF_ETIMER_EVAL(timer);
		tracerec->xor_us += RF_ETIMER_VAL_US(timer);
#endif
	}
	return (rf_GenericWakeupFunc(node, retcode));	/* call wake func
							 * explicitly since no
							 * I/O in this node */
}
/* this xor is used by the degraded-mode dag functions to recover lost
 * data.  the second-to-last parameter is the PDA for the failed
 * portion of the access.  the code here looks at this PDA and assumes
 * that the xor target buffer is equal in size to the number of
 * sectors in the failed PDA.  It then uses the other PDAs in the
 * parameter list to determine where within the target buffer the
 * corresponding data should be xored.  */
int
rf_RecoveryXorFunc(RF_DagNode_t *node)
{
	RF_Raid_t *raidPtr = (RF_Raid_t *) node->params[node->numParams - 1].p;
	RF_RaidLayout_t *layoutPtr = (RF_RaidLayout_t *) & raidPtr->Layout;
	RF_PhysDiskAddr_t *failedPDA = (RF_PhysDiskAddr_t *) node->params[node->numParams - 2].p;
	int     i, retcode = 0;
	RF_PhysDiskAddr_t *pda;
	int     suoffset, failedSUOffset = rf_StripeUnitOffset(layoutPtr, failedPDA->startSector);
	char   *srcbuf, *destbuf;
#if RF_ACC_TRACE > 0
	RF_AccTraceEntry_t *tracerec = node->dagHdr->tracerec;
	RF_Etimer_t timer;
#endif

	if (node->dagHdr->status == rf_enable) {
#if RF_ACC_TRACE > 0
		RF_ETIMER_START(timer);
#endif
		for (i = 0; i < node->numParams - 2; i += 2)
			if (node->params[i + 1].p != node->results[0]) {
				pda = (RF_PhysDiskAddr_t *) node->params[i].p;
				srcbuf = (char *) node->params[i + 1].p;
				suoffset = rf_StripeUnitOffset(layoutPtr, pda->startSector);
				destbuf = ((char *) node->results[0]) + rf_RaidAddressToByte(raidPtr, suoffset - failedSUOffset);
				retcode = rf_bxor(srcbuf, destbuf, rf_RaidAddressToByte(raidPtr, pda->numSector));
			}
#if RF_ACC_TRACE > 0
		RF_ETIMER_STOP(timer);
		RF_ETIMER_EVAL(timer);
		tracerec->xor_us += RF_ETIMER_VAL_US(timer);
#endif
	}
	return (rf_GenericWakeupFunc(node, retcode));
}
/*****************************************************************************
 * The next three functions are utilities used by the above
 * xor-execution functions.
 ****************************************************************************/


/*
 * this is just a glorified buffer xor.  targbuf points to a buffer
 * that is one full stripe unit in size.  srcbuf points to a buffer
 * that may be less than 1 SU, but never more.  When the access
 * described by pda is one SU in size (which by implication means it's
 * SU-aligned), all that happens is (targbuf) <- (srcbuf ^ targbuf).
 * When the access is less than one SU in size the XOR occurs on only
 * the portion of targbuf identified in the pda.  */

int
rf_XorIntoBuffer(RF_Raid_t *raidPtr, RF_PhysDiskAddr_t *pda,
		 char *srcbuf, char *targbuf)
{
	char   *targptr;
	int     sectPerSU = raidPtr->Layout.sectorsPerStripeUnit;
	int     SUOffset = pda->startSector % sectPerSU;
	int     length, retcode = 0;

	RF_ASSERT(pda->numSector <= sectPerSU);

	targptr = targbuf + rf_RaidAddressToByte(raidPtr, SUOffset);
	length = rf_RaidAddressToByte(raidPtr, pda->numSector);
	retcode = rf_bxor(srcbuf, targptr, length);
	return (retcode);
}
/* it really should be the case that the buffer pointers (returned by
 * malloc) are aligned to the natural word size of the machine, so
 * this is the only case we optimize for.  The length should always be
 * a multiple of the sector size, so there should be no problem with
 * leftover bytes at the end.  */
int
rf_bxor(char *src, char *dest, int len)
{
	unsigned mask = sizeof(long) - 1, retcode = 0;

	if (!(((unsigned long) src) & mask) &&
	    !(((unsigned long) dest) & mask) && !(len & mask)) {
		retcode = rf_longword_bxor((unsigned long *) src,
					   (unsigned long *) dest,
					   len >> RF_LONGSHIFT);
	} else {
		RF_ASSERT(0);
	}
	return (retcode);
}

/* When XORing in kernel mode, we need to map each user page to kernel
 * space before we can access it.  We don't want to assume anything
 * about which input buffers are in kernel/user space, nor about their
 * alignment, so in each loop we compute the maximum number of bytes
 * that we can xor without crossing any page boundaries, and do only
 * this many bytes before the next remap.
 *
 * len - is in longwords
 */
int
rf_longword_bxor(unsigned long *src, unsigned long *dest, int len)
{
	unsigned long *end = src + len;
	unsigned long d0, d1, d2, d3, s0, s1, s2, s3;	/* temps */
	unsigned long *pg_src, *pg_dest;   /* per-page source/dest pointers */
	int     longs_this_time;/* # longwords to xor in the current iteration */

	pg_src = src;
	pg_dest = dest;
	if (!pg_src || !pg_dest)
		return (EFAULT);

	while (len >= 4) {
		longs_this_time = RF_MIN(len, RF_MIN(RF_BLIP(pg_src), RF_BLIP(pg_dest)) >> RF_LONGSHIFT);	/* note len in longwords */
		src += longs_this_time;
		dest += longs_this_time;
		len -= longs_this_time;
		while (longs_this_time >= 4) {
			d0 = pg_dest[0];
			d1 = pg_dest[1];
			d2 = pg_dest[2];
			d3 = pg_dest[3];
			s0 = pg_src[0];
			s1 = pg_src[1];
			s2 = pg_src[2];
			s3 = pg_src[3];
			pg_dest[0] = d0 ^ s0;
			pg_dest[1] = d1 ^ s1;
			pg_dest[2] = d2 ^ s2;
			pg_dest[3] = d3 ^ s3;
			pg_src += 4;
			pg_dest += 4;
			longs_this_time -= 4;
		}
		while (longs_this_time > 0) {	/* cannot cross any page
						 * boundaries here */
			*pg_dest++ ^= *pg_src++;
			longs_this_time--;
		}

		/* either we're done, or we've reached a page boundary on one
		 * (or possibly both) of the pointers */
		if (len) {
			if (RF_PAGE_ALIGNED(src))
				pg_src = src;
			if (RF_PAGE_ALIGNED(dest))
				pg_dest = dest;
			if (!pg_src || !pg_dest)
				return (EFAULT);
		}
	}
	while (src < end) {
		*pg_dest++ ^= *pg_src++;
		src++;
		dest++;
		len--;
		if (RF_PAGE_ALIGNED(src))
			pg_src = src;
		if (RF_PAGE_ALIGNED(dest))
			pg_dest = dest;
	}
	RF_ASSERT(len == 0);
	return (0);
}

#if 0
/*
   dst = a ^ b ^ c;
   a may equal dst
   see comment above longword_bxor
   len is length in longwords
*/
int
rf_longword_bxor3(unsigned long *dst, unsigned long *a, unsigned long *b,
		  unsigned long *c, int len, void *bp)
{
	unsigned long a0, a1, a2, a3, b0, b1, b2, b3;
	unsigned long *pg_a, *pg_b, *pg_c, *pg_dst;	/* per-page source/dest
								 * pointers */
	int     longs_this_time;/* # longs to xor in the current iteration */
	char    dst_is_a = 0;

	pg_a = a;
	pg_b = b;
	pg_c = c;
	if (a == dst) {
		pg_dst = pg_a;
		dst_is_a = 1;
	} else {
		pg_dst = dst;
	}

	/* align dest to cache line.  Can't cross a pg boundary on dst here. */
	while ((((unsigned long) pg_dst) & 0x1f)) {
		*pg_dst++ = *pg_a++ ^ *pg_b++ ^ *pg_c++;
		dst++;
		a++;
		b++;
		c++;
		if (RF_PAGE_ALIGNED(a)) {
			pg_a = a;
			if (!pg_a)
				return (EFAULT);
		}
		if (RF_PAGE_ALIGNED(b)) {
			pg_b = a;
			if (!pg_b)
				return (EFAULT);
		}
		if (RF_PAGE_ALIGNED(c)) {
			pg_c = a;
			if (!pg_c)
				return (EFAULT);
		}
		len--;
	}

	while (len > 4) {
		longs_this_time = RF_MIN(len, RF_MIN(RF_BLIP(a), RF_MIN(RF_BLIP(b), RF_MIN(RF_BLIP(c), RF_BLIP(dst)))) >> RF_LONGSHIFT);
		a += longs_this_time;
		b += longs_this_time;
		c += longs_this_time;
		dst += longs_this_time;
		len -= longs_this_time;
		while (longs_this_time >= 4) {
			a0 = pg_a[0];
			longs_this_time -= 4;

			a1 = pg_a[1];
			a2 = pg_a[2];

			a3 = pg_a[3];
			pg_a += 4;

			b0 = pg_b[0];
			b1 = pg_b[1];

			b2 = pg_b[2];
			b3 = pg_b[3];
			/* start dual issue */
			a0 ^= b0;
			b0 = pg_c[0];

			pg_b += 4;
			a1 ^= b1;

			a2 ^= b2;
			a3 ^= b3;

			b1 = pg_c[1];
			a0 ^= b0;

			b2 = pg_c[2];
			a1 ^= b1;

			b3 = pg_c[3];
			a2 ^= b2;

			pg_dst[0] = a0;
			a3 ^= b3;
			pg_dst[1] = a1;
			pg_c += 4;
			pg_dst[2] = a2;
			pg_dst[3] = a3;
			pg_dst += 4;
		}
		while (longs_this_time > 0) {	/* cannot cross any page
						 * boundaries here */
			*pg_dst++ = *pg_a++ ^ *pg_b++ ^ *pg_c++;
			longs_this_time--;
		}

		if (len) {
			if (RF_PAGE_ALIGNED(a)) {
				pg_a = a;
				if (!pg_a)
					return (EFAULT);
				if (dst_is_a)
					pg_dst = pg_a;
			}
			if (RF_PAGE_ALIGNED(b)) {
				pg_b = b;
				if (!pg_b)
					return (EFAULT);
			}
			if (RF_PAGE_ALIGNED(c)) {
				pg_c = c;
				if (!pg_c)
					return (EFAULT);
			}
			if (!dst_is_a)
				if (RF_PAGE_ALIGNED(dst)) {
					pg_dst = dst;
					if (!pg_dst)
						return (EFAULT);
				}
		}
	}
	while (len) {
		*pg_dst++ = *pg_a++ ^ *pg_b++ ^ *pg_c++;
		dst++;
		a++;
		b++;
		c++;
		if (RF_PAGE_ALIGNED(a)) {
			pg_a = a;
			if (!pg_a)
				return (EFAULT);
			if (dst_is_a)
				pg_dst = pg_a;
		}
		if (RF_PAGE_ALIGNED(b)) {
			pg_b = b;
			if (!pg_b)
				return (EFAULT);
		}
		if (RF_PAGE_ALIGNED(c)) {
			pg_c = c;
			if (!pg_c)
				return (EFAULT);
		}
		if (!dst_is_a)
			if (RF_PAGE_ALIGNED(dst)) {
				pg_dst = dst;
				if (!pg_dst)
					return (EFAULT);
			}
		len--;
	}
	return (0);
}

int
rf_bxor3(unsigned char *dst, unsigned char *a, unsigned char *b,
	 unsigned char *c, unsigned long len, void *bp)
{
	RF_ASSERT(((RF_UL(dst) | RF_UL(a) | RF_UL(b) | RF_UL(c) | len) & 0x7) == 0);

	return (rf_longword_bxor3((unsigned long *) dst, (unsigned long *) a,
		(unsigned long *) b, (unsigned long *) c, len >> RF_LONGSHIFT, bp));
}
#endif